introduction_to_manual_feature_engineering_part_one.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "9d7057b0f4ce0d2bb50bfea37c29ce0927cdf53f"
   },
   "source": [
    "# Introduction: Manual Feature Engineering\n",
    "\n",
    "If you are new to this competition, I highly suggest checking out [this notebook](https://www.kaggle.com/willkoehrsen/start-here-a-gentle-introduction/) to get started.\n",
    "\n",
    "In this notebook, we will explore making features by hand for the Home Credit Default Risk competition. In an earlier notebook, we used only the `application` data in order to build a model. The best model we made from this data achieved a score on the leaderboard around 0.74. In order to better this score, we will have to include more information from the other dataframes. Here, we will look at using information from the `bureau` and `bureau_balance` data. The definitions of these data files are:\n",
    "\n",
    "* bureau: information about client's previous loans with other financial institutions reported to Home Credit. Each previous loan has its own row.\n",
    "* bureau_balance: monthly information about the previous loans. Each month has its own row.\n",
    "\n",
    "Manual feature engineering can be a tedious process (which is why we use automated feature engineering with featuretools!) and often relies on domain expertise. Since I have limited domain knowledge of loans and what makes a person likely to default, I will instead concentrate of getting as much info as possible into the final training dataframe. The idea is that the model will then pick up on which features are important rather than us having to decide that. Basically, our approach is to make as many features as possible and then give them all to the model to use! Later, we can perform feature reduction using the feature importances from the model or other techniques such as PCA. \n",
    "\n",
    "The process of manual feature engineering will involve plenty of Pandas code, a little patience, and a lot of great practice manipulation data. Even though automated feature engineering tools are starting to be made available, feature engineering will still have to be done using plenty of data wrangling for a little while longer. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
    "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# pandas and numpy for data manipulation\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "# matplotlib and seaborn for plotting\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "# Suppress warnings from pandas\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "\n",
    "plt.style.use('fivethirtyeight')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "b4e0980ff7f4d8d9f0661b5d5ebf07e66d304222"
   },
   "source": [
    "## Example: Counts of a client's previous loans\n",
    "\n",
    "To illustrate the general process of manual feature engineering, we will first simply get the count of a client's previous loans at other financial institutions. This requires a number of Pandas operations we will make heavy use of throughout the notebook:\n",
    "\n",
    "* `groupby`: group a dataframe by a column. In this case we will group by the unique client, the `SK_ID_CURR` column\n",
    "* `agg`: perform a calculation on the grouped data such as taking the mean of columns. We can either call the function directly (`grouped_df.mean()`) or use the `agg` function together with a list of transforms (`grouped_df.agg([mean, max, min, sum])`)\n",
    "* `merge`: match the aggregated statistics to the appropriate client. We need to merge the original training data with the calculated stats on the `SK_ID_CURR` column which will insert `NaN` in any cell for which the client does not have the corresponding statistic\n",
    "\n",
    "We also use the (`rename`) function quite a bit specifying the columns to be renamed as a dictionary. This is useful in order to keep track of the new variables we create.\n",
    "\n",
    "This might seem like a lot, which is why we'll eventually write a function to do this process for us. Let's take a look at implementing this by hand first. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "_cell_guid": "79c7e3d0-c299-4dcb-8224-4455121ee9b0",
    "_uuid": "d629ff2d2480ee46fbb7e2d37f6b5fab8052498a",
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th>SK_ID_BUREAU</th>\n",
       "      <th>CREDIT_ACTIVE</th>\n",
       "      <th>CREDIT_CURRENCY</th>\n",
       "      <th>DAYS_CREDIT</th>\n",
       "      <th>CREDIT_DAY_OVERDUE</th>\n",
       "      <th>DAYS_CREDIT_ENDDATE</th>\n",
       "      <th>DAYS_ENDDATE_FACT</th>\n",
       "      <th>AMT_CREDIT_MAX_OVERDUE</th>\n",
       "      <th>CNT_CREDIT_PROLONG</th>\n",
       "      <th>AMT_CREDIT_SUM</th>\n",
       "      <th>AMT_CREDIT_SUM_DEBT</th>\n",
       "      <th>AMT_CREDIT_SUM_LIMIT</th>\n",
       "      <th>AMT_CREDIT_SUM_OVERDUE</th>\n",
       "      <th>CREDIT_TYPE</th>\n",
       "      <th>DAYS_CREDIT_UPDATE</th>\n",
       "      <th>AMT_ANNUITY</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>215354</td>\n",
       "      <td>5714462</td>\n",
       "      <td>Closed</td>\n",
       "      <td>currency 1</td>\n",
       "      <td>-497</td>\n",
       "      <td>0</td>\n",
       "      <td>-153.0</td>\n",
       "      <td>-153.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>91323.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>Consumer credit</td>\n",
       "      <td>-131</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>215354</td>\n",
       "      <td>5714463</td>\n",
       "      <td>Active</td>\n",
       "      <td>currency 1</td>\n",
       "      <td>-208</td>\n",
       "      <td>0</td>\n",
       "      <td>1075.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>225000.0</td>\n",
       "      <td>171342.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>Credit card</td>\n",
       "      <td>-20</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>215354</td>\n",
       "      <td>5714464</td>\n",
       "      <td>Active</td>\n",
       "      <td>currency 1</td>\n",
       "      <td>-203</td>\n",
       "      <td>0</td>\n",
       "      <td>528.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>464323.5</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>Consumer credit</td>\n",
       "      <td>-16</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>215354</td>\n",
       "      <td>5714465</td>\n",
       "      <td>Active</td>\n",
       "      <td>currency 1</td>\n",
       "      <td>-203</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>90000.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>Credit card</td>\n",
       "      <td>-16</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>215354</td>\n",
       "      <td>5714466</td>\n",
       "      <td>Active</td>\n",
       "      <td>currency 1</td>\n",
       "      <td>-629</td>\n",
       "      <td>0</td>\n",
       "      <td>1197.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>77674.5</td>\n",
       "      <td>0</td>\n",
       "      <td>2700000.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>Consumer credit</td>\n",
       "      <td>-21</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_CURR  SK_ID_BUREAU CREDIT_ACTIVE CREDIT_CURRENCY  DAYS_CREDIT  \\\n",
       "0      215354       5714462        Closed      currency 1         -497   \n",
       "1      215354       5714463        Active      currency 1         -208   \n",
       "2      215354       5714464        Active      currency 1         -203   \n",
       "3      215354       5714465        Active      currency 1         -203   \n",
       "4      215354       5714466        Active      currency 1         -629   \n",
       "\n",
       "   CREDIT_DAY_OVERDUE  DAYS_CREDIT_ENDDATE  DAYS_ENDDATE_FACT  \\\n",
       "0                   0               -153.0             -153.0   \n",
       "1                   0               1075.0                NaN   \n",
       "2                   0                528.0                NaN   \n",
       "3                   0                  NaN                NaN   \n",
       "4                   0               1197.0                NaN   \n",
       "\n",
       "   AMT_CREDIT_MAX_OVERDUE  CNT_CREDIT_PROLONG  AMT_CREDIT_SUM  \\\n",
       "0                     NaN                   0         91323.0   \n",
       "1                     NaN                   0        225000.0   \n",
       "2                     NaN                   0        464323.5   \n",
       "3                     NaN                   0         90000.0   \n",
       "4                 77674.5                   0       2700000.0   \n",
       "\n",
       "   AMT_CREDIT_SUM_DEBT  AMT_CREDIT_SUM_LIMIT  AMT_CREDIT_SUM_OVERDUE  \\\n",
       "0                  0.0                   NaN                     0.0   \n",
       "1             171342.0                   NaN                     0.0   \n",
       "2                  NaN                   NaN                     0.0   \n",
       "3                  NaN                   NaN                     0.0   \n",
       "4                  NaN                   NaN                     0.0   \n",
       "\n",
       "       CREDIT_TYPE  DAYS_CREDIT_UPDATE  AMT_ANNUITY  \n",
       "0  Consumer credit                -131          NaN  \n",
       "1      Credit card                 -20          NaN  \n",
       "2  Consumer credit                 -16          NaN  \n",
       "3      Credit card                 -16          NaN  \n",
       "4  Consumer credit                 -21          NaN  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Read in bureau\n",
    "bureau = pd.read_csv('../input/bureau.csv')\n",
    "bureau.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "_uuid": "6665d87bd3a157c20fb5a383322aa153005cad2b",
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th>previous_loan_counts</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>100001</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>100002</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>100003</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>100004</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>100005</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_CURR  previous_loan_counts\n",
       "0      100001                     7\n",
       "1      100002                     8\n",
       "2      100003                     4\n",
       "3      100004                     2\n",
       "4      100005                     3"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Groupby the client id (SK_ID_CURR), count the number of previous loans, and rename the column\n",
    "previous_loan_counts = bureau.groupby('SK_ID_CURR', as_index=False)['SK_ID_BUREAU'].count().rename(columns = {'SK_ID_BUREAU': 'previous_loan_counts'})\n",
    "previous_loan_counts.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "_uuid": "1adcc4acb891adae8646211db629fb263660c2bb",
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th>TARGET</th>\n",
       "      <th>NAME_CONTRACT_TYPE</th>\n",
       "      <th>CODE_GENDER</th>\n",
       "      <th>FLAG_OWN_CAR</th>\n",
       "      <th>FLAG_OWN_REALTY</th>\n",
       "      <th>CNT_CHILDREN</th>\n",
       "      <th>AMT_INCOME_TOTAL</th>\n",
       "      <th>AMT_CREDIT</th>\n",
       "      <th>AMT_ANNUITY</th>\n",
       "      <th>...</th>\n",
       "      <th>FLAG_DOCUMENT_19</th>\n",
       "      <th>FLAG_DOCUMENT_20</th>\n",
       "      <th>FLAG_DOCUMENT_21</th>\n",
       "      <th>AMT_REQ_CREDIT_BUREAU_HOUR</th>\n",
       "      <th>AMT_REQ_CREDIT_BUREAU_DAY</th>\n",
       "      <th>AMT_REQ_CREDIT_BUREAU_WEEK</th>\n",
       "      <th>AMT_REQ_CREDIT_BUREAU_MON</th>\n",
       "      <th>AMT_REQ_CREDIT_BUREAU_QRT</th>\n",
       "      <th>AMT_REQ_CREDIT_BUREAU_YEAR</th>\n",
       "      <th>previous_loan_counts</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>100002</td>\n",
       "      <td>1</td>\n",
       "      <td>Cash loans</td>\n",
       "      <td>M</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>202500.0</td>\n",
       "      <td>406597.5</td>\n",
       "      <td>24700.5</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>8.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>100003</td>\n",
       "      <td>0</td>\n",
       "      <td>Cash loans</td>\n",
       "      <td>F</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0</td>\n",
       "      <td>270000.0</td>\n",
       "      <td>1293502.5</td>\n",
       "      <td>35698.5</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>100004</td>\n",
       "      <td>0</td>\n",
       "      <td>Revolving loans</td>\n",
       "      <td>M</td>\n",
       "      <td>Y</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>67500.0</td>\n",
       "      <td>135000.0</td>\n",
       "      <td>6750.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>100006</td>\n",
       "      <td>0</td>\n",
       "      <td>Cash loans</td>\n",
       "      <td>F</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>135000.0</td>\n",
       "      <td>312682.5</td>\n",
       "      <td>29686.5</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>100007</td>\n",
       "      <td>0</td>\n",
       "      <td>Cash loans</td>\n",
       "      <td>M</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>121500.0</td>\n",
       "      <td>513000.0</td>\n",
       "      <td>21865.5</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 123 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_CURR  TARGET NAME_CONTRACT_TYPE CODE_GENDER FLAG_OWN_CAR  \\\n",
       "0      100002       1         Cash loans           M            N   \n",
       "1      100003       0         Cash loans           F            N   \n",
       "2      100004       0    Revolving loans           M            Y   \n",
       "3      100006       0         Cash loans           F            N   \n",
       "4      100007       0         Cash loans           M            N   \n",
       "\n",
       "  FLAG_OWN_REALTY  CNT_CHILDREN  AMT_INCOME_TOTAL  AMT_CREDIT  AMT_ANNUITY  \\\n",
       "0               Y             0          202500.0    406597.5      24700.5   \n",
       "1               N             0          270000.0   1293502.5      35698.5   \n",
       "2               Y             0           67500.0    135000.0       6750.0   \n",
       "3               Y             0          135000.0    312682.5      29686.5   \n",
       "4               Y             0          121500.0    513000.0      21865.5   \n",
       "\n",
       "           ...           FLAG_DOCUMENT_19 FLAG_DOCUMENT_20 FLAG_DOCUMENT_21  \\\n",
       "0          ...                          0                0                0   \n",
       "1          ...                          0                0                0   \n",
       "2          ...                          0                0                0   \n",
       "3          ...                          0                0                0   \n",
       "4          ...                          0                0                0   \n",
       "\n",
       "  AMT_REQ_CREDIT_BUREAU_HOUR AMT_REQ_CREDIT_BUREAU_DAY  \\\n",
       "0                        0.0                       0.0   \n",
       "1                        0.0                       0.0   \n",
       "2                        0.0                       0.0   \n",
       "3                        NaN                       NaN   \n",
       "4                        0.0                       0.0   \n",
       "\n",
       "  AMT_REQ_CREDIT_BUREAU_WEEK  AMT_REQ_CREDIT_BUREAU_MON  \\\n",
       "0                        0.0                        0.0   \n",
       "1                        0.0                        0.0   \n",
       "2                        0.0                        0.0   \n",
       "3                        NaN                        NaN   \n",
       "4                        0.0                        0.0   \n",
       "\n",
       "   AMT_REQ_CREDIT_BUREAU_QRT  AMT_REQ_CREDIT_BUREAU_YEAR  previous_loan_counts  \n",
       "0                        0.0                         1.0                   8.0  \n",
       "1                        0.0                         0.0                   4.0  \n",
       "2                        0.0                         0.0                   2.0  \n",
       "3                        NaN                         NaN                   0.0  \n",
       "4                        0.0                         0.0                   1.0  \n",
       "\n",
       "[5 rows x 123 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Join to the training dataframe\n",
    "train = pd.read_csv('../input/application_train.csv')\n",
    "train = train.merge(previous_loan_counts, on = 'SK_ID_CURR', how = 'left')\n",
    "\n",
    "# Fill the missing values with 0 \n",
    "train['previous_loan_counts'] = train['previous_loan_counts'].fillna(0)\n",
    "train.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "53d83da3d8a28541c2dd49be8047e110034c10e6"
   },
   "source": [
    "Scroll all the way to the right to see the new column. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "173b8548125c8a344f67d397d7c24af223af7254"
   },
   "source": [
    "## Assessing Usefulness of New Variable with r value\n",
    "\n",
    "To determine if the new variable is useful, we can calculate the Pearson Correlation Coefficient (r-value) between this variable and the target. This measures the strength of a linear relationship between two variables and ranges from -1 (perfectly negatively linear) to +1 (perfectly positively linear). The r-value is not best measure of the \"usefulness\" of a new variable, but it can give a first approximation of whether a variable will be helpful to a machine learning model. The larger the r-value of a variable with respect to the target, the more a change in this variable is likely to affect the value of the target. Therefore, we look for the variables with the greatest absolute value r-value relative to the target.\n",
    "\n",
    "We can also visually inspect a relationship with the target using the Kernel Density Estimate (KDE) plot. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "f5bc72f178a44d45399f620ee43483df2482c02f"
   },
   "source": [
    "### Kernel Density Estimate Plots\n",
    "\n",
    "The kernel density estimate plot shows the distribution of a single variable (think of it as a smoothed histogram). To see the different in distributions dependent on the value of a categorical variable, we can color the distributions differently according to the category. For example, we can show the kernel density estimate of the `previous_loan_count` colored by whether the `TARGET` = 1 or 0. The resulting KDE will show any significant differences in the distribution of the variable between people who did not repay their loan (`TARGET == 1`) and the people who did (`TARGET == 0`). This can serve as an indicator of whether a variable will be 'relevant' to a machine learning model. \n",
    "\n",
    "We will put this plotting functionality in a function to re-use for any variable. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "_uuid": "4a5fab58ab8327c7f53361b584603d78df96c7a4",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Plots the disribution of a variable colored by value of the target\n",
    "def kde_target(var_name, df):\n",
    "    \n",
    "    # Calculate the correlation coefficient between the new variable and the target\n",
    "    corr = df['TARGET'].corr(df[var_name])\n",
    "    \n",
    "    # Calculate medians for repaid vs not repaid\n",
    "    avg_repaid = df.ix[df['TARGET'] == 0, var_name].median()\n",
    "    avg_not_repaid = df.ix[df['TARGET'] == 1, var_name].median()\n",
    "    \n",
    "    plt.figure(figsize = (12, 6))\n",
    "    \n",
    "    # Plot the distribution for target == 0 and target == 1\n",
    "    sns.kdeplot(df.ix[df['TARGET'] == 0, var_name], label = 'TARGET == 0')\n",
    "    sns.kdeplot(df.ix[df['TARGET'] == 1, var_name], label = 'TARGET == 1')\n",
    "    \n",
    "    # label the plot\n",
    "    plt.xlabel(var_name); plt.ylabel('Density'); plt.title('%s Distribution' % var_name)\n",
    "    plt.legend();\n",
    "    \n",
    "    # print out the correlation\n",
    "    print('The correlation between %s and the TARGET is %0.4f' % (var_name, corr))\n",
    "    # Print out average values\n",
    "    print('Median value for loan that was not repaid = %0.4f' % avg_not_repaid)\n",
    "    print('Median value for loan that was repaid =     %0.4f' % avg_repaid)\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "783e2b23b473f5f61d17ba5b5b64349d3a1d60cf"
   },
   "source": [
    "We can test this function using the `EXT_SOURCE_3` variable which we [found to be one of the most important variables ](https://www.kaggle.com/willkoehrsen/start-here-a-gentle-introduction) according to a Random Forest and Gradient Boosting Machine. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "_uuid": "f9e51eb7f02c5612edef3c950ae89da3b8ececf4",
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The correlation between EXT_SOURCE_3 and the TARGET is -0.1789\n",
      "Median value for loan that was not repaid = 0.3791\n",
      "Median value for loan that was repaid =     0.5460\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "kde_target('EXT_SOURCE_3', train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "c4598cb54ad22b04646683832e2e4c4bc98ff8e9"
   },
   "source": [
    "Now for the new variable we just made, the number of previous loans at other institutions."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "_uuid": "bbae3d4e85f106c1e7b457ad71e1549c0e7b1267",
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The correlation between previous_loan_counts and the TARGET is -0.0100\n",
      "Median value for loan that was not repaid = 3.0000\n",
      "Median value for loan that was repaid =     4.0000\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "kde_target('previous_loan_counts', train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "a85c4e3f8ccb54bb4dbbd85a414744d301916fd8"
   },
   "source": [
    "From this it's difficult to tell if this variable will be important. The correlation coefficient is extremely weak and there is almost no noticeable difference in the distributions. \n",
    "\n",
    "Let's move on to make a few more variables from the bureau dataframe. We will take the mean, min, and max of every numeric column in the bureau dataframe."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "11f1b0ab2146ec95a50055d2788b5f05c43e0afb"
   },
   "source": [
    "## Aggregating Numeric Columns\n",
    "\n",
    "To account for the numeric information in the `bureau` dataframe, we can compute statistics for all the numeric columns. To do so, we `groupby` the client id, `agg` the grouped dataframe, and merge the result back into the training data. The `agg` function will only calculate the values for the numeric columns where the operation is considered valid. We will stick to using `'mean', 'max', 'min', 'sum'` but any function can be passed in here. We can even write our own function and use it in an `agg` call. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "_uuid": "582885d850fa1316ce869b856b9d886ad542b04b",
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th colspan=\"4\" halign=\"left\">DAYS_CREDIT</th>\n",
       "      <th colspan=\"4\" halign=\"left\">CREDIT_DAY_OVERDUE</th>\n",
       "      <th>DAYS_CREDIT_ENDDATE</th>\n",
       "      <th>...</th>\n",
       "      <th colspan=\"2\" halign=\"left\">AMT_CREDIT_SUM_OVERDUE</th>\n",
       "      <th colspan=\"4\" halign=\"left\">DAYS_CREDIT_UPDATE</th>\n",
       "      <th colspan=\"4\" halign=\"left\">AMT_ANNUITY</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>mean</th>\n",
       "      <th>max</th>\n",
       "      <th>min</th>\n",
       "      <th>sum</th>\n",
       "      <th>mean</th>\n",
       "      <th>max</th>\n",
       "      <th>min</th>\n",
       "      <th>sum</th>\n",
       "      <th>mean</th>\n",
       "      <th>...</th>\n",
       "      <th>min</th>\n",
       "      <th>sum</th>\n",
       "      <th>mean</th>\n",
       "      <th>max</th>\n",
       "      <th>min</th>\n",
       "      <th>sum</th>\n",
       "      <th>mean</th>\n",
       "      <th>max</th>\n",
       "      <th>min</th>\n",
       "      <th>sum</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>100001</td>\n",
       "      <td>-735.000000</td>\n",
       "      <td>-49</td>\n",
       "      <td>-1572</td>\n",
       "      <td>-5145</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>82.428571</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-93.142857</td>\n",
       "      <td>-6</td>\n",
       "      <td>-155</td>\n",
       "      <td>-652</td>\n",
       "      <td>3545.357143</td>\n",
       "      <td>10822.5</td>\n",
       "      <td>0.0</td>\n",
       "      <td>24817.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>100002</td>\n",
       "      <td>-874.000000</td>\n",
       "      <td>-103</td>\n",
       "      <td>-1437</td>\n",
       "      <td>-6992</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-349.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-499.875000</td>\n",
       "      <td>-7</td>\n",
       "      <td>-1185</td>\n",
       "      <td>-3999</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>100003</td>\n",
       "      <td>-1400.750000</td>\n",
       "      <td>-606</td>\n",
       "      <td>-2586</td>\n",
       "      <td>-5603</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-544.500000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-816.000000</td>\n",
       "      <td>-43</td>\n",
       "      <td>-2131</td>\n",
       "      <td>-3264</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>100004</td>\n",
       "      <td>-867.000000</td>\n",
       "      <td>-408</td>\n",
       "      <td>-1326</td>\n",
       "      <td>-1734</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-488.500000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-532.000000</td>\n",
       "      <td>-382</td>\n",
       "      <td>-682</td>\n",
       "      <td>-1064</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>100005</td>\n",
       "      <td>-190.666667</td>\n",
       "      <td>-62</td>\n",
       "      <td>-373</td>\n",
       "      <td>-572</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>439.333333</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-54.333333</td>\n",
       "      <td>-11</td>\n",
       "      <td>-121</td>\n",
       "      <td>-163</td>\n",
       "      <td>1420.500000</td>\n",
       "      <td>4261.5</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4261.5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 49 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "  SK_ID_CURR  DAYS_CREDIT                  CREDIT_DAY_OVERDUE              \\\n",
       "                     mean  max   min   sum               mean max min sum   \n",
       "0     100001  -735.000000  -49 -1572 -5145                0.0   0   0   0   \n",
       "1     100002  -874.000000 -103 -1437 -6992                0.0   0   0   0   \n",
       "2     100003 -1400.750000 -606 -2586 -5603                0.0   0   0   0   \n",
       "3     100004  -867.000000 -408 -1326 -1734                0.0   0   0   0   \n",
       "4     100005  -190.666667  -62  -373  -572                0.0   0   0   0   \n",
       "\n",
       "  DAYS_CREDIT_ENDDATE   ...    AMT_CREDIT_SUM_OVERDUE      DAYS_CREDIT_UPDATE  \\\n",
       "                 mean   ...                       min  sum               mean   \n",
       "0           82.428571   ...                       0.0  0.0         -93.142857   \n",
       "1         -349.000000   ...                       0.0  0.0        -499.875000   \n",
       "2         -544.500000   ...                       0.0  0.0        -816.000000   \n",
       "3         -488.500000   ...                       0.0  0.0        -532.000000   \n",
       "4          439.333333   ...                       0.0  0.0         -54.333333   \n",
       "\n",
       "                    AMT_ANNUITY                         \n",
       "   max   min   sum         mean      max  min      sum  \n",
       "0   -6  -155  -652  3545.357143  10822.5  0.0  24817.5  \n",
       "1   -7 -1185 -3999     0.000000      0.0  0.0      0.0  \n",
       "2  -43 -2131 -3264          NaN      NaN  NaN      0.0  \n",
       "3 -382  -682 -1064          NaN      NaN  NaN      0.0  \n",
       "4  -11  -121  -163  1420.500000   4261.5  0.0   4261.5  \n",
       "\n",
       "[5 rows x 49 columns]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Group by the client id, calculate aggregation statistics\n",
    "bureau_agg = bureau.drop(columns = ['SK_ID_BUREAU']).groupby('SK_ID_CURR', as_index = False).agg(['mean', 'max', 'min', 'sum']).reset_index()\n",
    "bureau_agg.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "812f7b1e0d6016768d9276da790b0986181f9fdd"
   },
   "source": [
    "We need to create new names for each of these columns. The following code makes new names by appending the stat to the name. Here we have to deal with the fact that the dataframe has a multi-level index. I find these confusing and hard to work with, so I try to reduce to a single level index as quickly as possible."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "_uuid": "01acafad68d285c875018ae916c002e9dad1a2fa",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# List of column names\n",
    "columns = ['SK_ID_CURR']\n",
    "\n",
    "# Iterate through the variables names\n",
    "for var in bureau_agg.columns.levels[0]:\n",
    "    # Skip the id name\n",
    "    if var != 'SK_ID_CURR':\n",
    "        \n",
    "        # Iterate through the stat names\n",
    "        for stat in bureau_agg.columns.levels[1][:-1]:\n",
    "            # Make a new column name for the variable and stat\n",
    "            columns.append('bureau_%s_%s' % (var, stat))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "_uuid": "4f34abd311ffcb34e9feb0786998e478ffd90766",
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th>bureau_DAYS_CREDIT_mean</th>\n",
       "      <th>bureau_DAYS_CREDIT_max</th>\n",
       "      <th>bureau_DAYS_CREDIT_min</th>\n",
       "      <th>bureau_DAYS_CREDIT_sum</th>\n",
       "      <th>bureau_CREDIT_DAY_OVERDUE_mean</th>\n",
       "      <th>bureau_CREDIT_DAY_OVERDUE_max</th>\n",
       "      <th>bureau_CREDIT_DAY_OVERDUE_min</th>\n",
       "      <th>bureau_CREDIT_DAY_OVERDUE_sum</th>\n",
       "      <th>bureau_DAYS_CREDIT_ENDDATE_mean</th>\n",
       "      <th>...</th>\n",
       "      <th>bureau_AMT_CREDIT_SUM_OVERDUE_min</th>\n",
       "      <th>bureau_AMT_CREDIT_SUM_OVERDUE_sum</th>\n",
       "      <th>bureau_DAYS_CREDIT_UPDATE_mean</th>\n",
       "      <th>bureau_DAYS_CREDIT_UPDATE_max</th>\n",
       "      <th>bureau_DAYS_CREDIT_UPDATE_min</th>\n",
       "      <th>bureau_DAYS_CREDIT_UPDATE_sum</th>\n",
       "      <th>bureau_AMT_ANNUITY_mean</th>\n",
       "      <th>bureau_AMT_ANNUITY_max</th>\n",
       "      <th>bureau_AMT_ANNUITY_min</th>\n",
       "      <th>bureau_AMT_ANNUITY_sum</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>100001</td>\n",
       "      <td>-735.000000</td>\n",
       "      <td>-49</td>\n",
       "      <td>-1572</td>\n",
       "      <td>-5145</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>82.428571</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-93.142857</td>\n",
       "      <td>-6</td>\n",
       "      <td>-155</td>\n",
       "      <td>-652</td>\n",
       "      <td>3545.357143</td>\n",
       "      <td>10822.5</td>\n",
       "      <td>0.0</td>\n",
       "      <td>24817.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>100002</td>\n",
       "      <td>-874.000000</td>\n",
       "      <td>-103</td>\n",
       "      <td>-1437</td>\n",
       "      <td>-6992</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-349.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-499.875000</td>\n",
       "      <td>-7</td>\n",
       "      <td>-1185</td>\n",
       "      <td>-3999</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>100003</td>\n",
       "      <td>-1400.750000</td>\n",
       "      <td>-606</td>\n",
       "      <td>-2586</td>\n",
       "      <td>-5603</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-544.500000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-816.000000</td>\n",
       "      <td>-43</td>\n",
       "      <td>-2131</td>\n",
       "      <td>-3264</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>100004</td>\n",
       "      <td>-867.000000</td>\n",
       "      <td>-408</td>\n",
       "      <td>-1326</td>\n",
       "      <td>-1734</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-488.500000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-532.000000</td>\n",
       "      <td>-382</td>\n",
       "      <td>-682</td>\n",
       "      <td>-1064</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>100005</td>\n",
       "      <td>-190.666667</td>\n",
       "      <td>-62</td>\n",
       "      <td>-373</td>\n",
       "      <td>-572</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>439.333333</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-54.333333</td>\n",
       "      <td>-11</td>\n",
       "      <td>-121</td>\n",
       "      <td>-163</td>\n",
       "      <td>1420.500000</td>\n",
       "      <td>4261.5</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4261.5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 49 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_CURR  bureau_DAYS_CREDIT_mean  bureau_DAYS_CREDIT_max  \\\n",
       "0      100001              -735.000000                     -49   \n",
       "1      100002              -874.000000                    -103   \n",
       "2      100003             -1400.750000                    -606   \n",
       "3      100004              -867.000000                    -408   \n",
       "4      100005              -190.666667                     -62   \n",
       "\n",
       "   bureau_DAYS_CREDIT_min  bureau_DAYS_CREDIT_sum  \\\n",
       "0                   -1572                   -5145   \n",
       "1                   -1437                   -6992   \n",
       "2                   -2586                   -5603   \n",
       "3                   -1326                   -1734   \n",
       "4                    -373                    -572   \n",
       "\n",
       "   bureau_CREDIT_DAY_OVERDUE_mean  bureau_CREDIT_DAY_OVERDUE_max  \\\n",
       "0                             0.0                              0   \n",
       "1                             0.0                              0   \n",
       "2                             0.0                              0   \n",
       "3                             0.0                              0   \n",
       "4                             0.0                              0   \n",
       "\n",
       "   bureau_CREDIT_DAY_OVERDUE_min  bureau_CREDIT_DAY_OVERDUE_sum  \\\n",
       "0                              0                              0   \n",
       "1                              0                              0   \n",
       "2                              0                              0   \n",
       "3                              0                              0   \n",
       "4                              0                              0   \n",
       "\n",
       "   bureau_DAYS_CREDIT_ENDDATE_mean           ...            \\\n",
       "0                        82.428571           ...             \n",
       "1                      -349.000000           ...             \n",
       "2                      -544.500000           ...             \n",
       "3                      -488.500000           ...             \n",
       "4                       439.333333           ...             \n",
       "\n",
       "   bureau_AMT_CREDIT_SUM_OVERDUE_min  bureau_AMT_CREDIT_SUM_OVERDUE_sum  \\\n",
       "0                                0.0                                0.0   \n",
       "1                                0.0                                0.0   \n",
       "2                                0.0                                0.0   \n",
       "3                                0.0                                0.0   \n",
       "4                                0.0                                0.0   \n",
       "\n",
       "   bureau_DAYS_CREDIT_UPDATE_mean  bureau_DAYS_CREDIT_UPDATE_max  \\\n",
       "0                      -93.142857                             -6   \n",
       "1                     -499.875000                             -7   \n",
       "2                     -816.000000                            -43   \n",
       "3                     -532.000000                           -382   \n",
       "4                      -54.333333                            -11   \n",
       "\n",
       "   bureau_DAYS_CREDIT_UPDATE_min  bureau_DAYS_CREDIT_UPDATE_sum  \\\n",
       "0                           -155                           -652   \n",
       "1                          -1185                          -3999   \n",
       "2                          -2131                          -3264   \n",
       "3                           -682                          -1064   \n",
       "4                           -121                           -163   \n",
       "\n",
       "   bureau_AMT_ANNUITY_mean  bureau_AMT_ANNUITY_max  bureau_AMT_ANNUITY_min  \\\n",
       "0              3545.357143                 10822.5                     0.0   \n",
       "1                 0.000000                     0.0                     0.0   \n",
       "2                      NaN                     NaN                     NaN   \n",
       "3                      NaN                     NaN                     NaN   \n",
       "4              1420.500000                  4261.5                     0.0   \n",
       "\n",
       "   bureau_AMT_ANNUITY_sum  \n",
       "0                 24817.5  \n",
       "1                     0.0  \n",
       "2                     0.0  \n",
       "3                     0.0  \n",
       "4                  4261.5  \n",
       "\n",
       "[5 rows x 49 columns]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Assign the list of columns names as the dataframe column names\n",
    "bureau_agg.columns = columns\n",
    "bureau_agg.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "13c6a8b69a3bdbd1bf2d420d984f78400fc969d3"
   },
   "source": [
    "Now we simply merge with the training data as we did before."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "_uuid": "2da1e6aa5651f4567813896d79d2e087fd5e7749",
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th>TARGET</th>\n",
       "      <th>NAME_CONTRACT_TYPE</th>\n",
       "      <th>CODE_GENDER</th>\n",
       "      <th>FLAG_OWN_CAR</th>\n",
       "      <th>FLAG_OWN_REALTY</th>\n",
       "      <th>CNT_CHILDREN</th>\n",
       "      <th>AMT_INCOME_TOTAL</th>\n",
       "      <th>AMT_CREDIT</th>\n",
       "      <th>AMT_ANNUITY</th>\n",
       "      <th>...</th>\n",
       "      <th>bureau_AMT_CREDIT_SUM_OVERDUE_min</th>\n",
       "      <th>bureau_AMT_CREDIT_SUM_OVERDUE_sum</th>\n",
       "      <th>bureau_DAYS_CREDIT_UPDATE_mean</th>\n",
       "      <th>bureau_DAYS_CREDIT_UPDATE_max</th>\n",
       "      <th>bureau_DAYS_CREDIT_UPDATE_min</th>\n",
       "      <th>bureau_DAYS_CREDIT_UPDATE_sum</th>\n",
       "      <th>bureau_AMT_ANNUITY_mean</th>\n",
       "      <th>bureau_AMT_ANNUITY_max</th>\n",
       "      <th>bureau_AMT_ANNUITY_min</th>\n",
       "      <th>bureau_AMT_ANNUITY_sum</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>100002</td>\n",
       "      <td>1</td>\n",
       "      <td>Cash loans</td>\n",
       "      <td>M</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>202500.0</td>\n",
       "      <td>406597.5</td>\n",
       "      <td>24700.5</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-499.875</td>\n",
       "      <td>-7.0</td>\n",
       "      <td>-1185.0</td>\n",
       "      <td>-3999.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>100003</td>\n",
       "      <td>0</td>\n",
       "      <td>Cash loans</td>\n",
       "      <td>F</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0</td>\n",
       "      <td>270000.0</td>\n",
       "      <td>1293502.5</td>\n",
       "      <td>35698.5</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-816.000</td>\n",
       "      <td>-43.0</td>\n",
       "      <td>-2131.0</td>\n",
       "      <td>-3264.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>100004</td>\n",
       "      <td>0</td>\n",
       "      <td>Revolving loans</td>\n",
       "      <td>M</td>\n",
       "      <td>Y</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>67500.0</td>\n",
       "      <td>135000.0</td>\n",
       "      <td>6750.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-532.000</td>\n",
       "      <td>-382.0</td>\n",
       "      <td>-682.0</td>\n",
       "      <td>-1064.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>100006</td>\n",
       "      <td>0</td>\n",
       "      <td>Cash loans</td>\n",
       "      <td>F</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>135000.0</td>\n",
       "      <td>312682.5</td>\n",
       "      <td>29686.5</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>100007</td>\n",
       "      <td>0</td>\n",
       "      <td>Cash loans</td>\n",
       "      <td>M</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>121500.0</td>\n",
       "      <td>513000.0</td>\n",
       "      <td>21865.5</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-783.000</td>\n",
       "      <td>-783.0</td>\n",
       "      <td>-783.0</td>\n",
       "      <td>-783.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 171 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_CURR  TARGET NAME_CONTRACT_TYPE CODE_GENDER FLAG_OWN_CAR  \\\n",
       "0      100002       1         Cash loans           M            N   \n",
       "1      100003       0         Cash loans           F            N   \n",
       "2      100004       0    Revolving loans           M            Y   \n",
       "3      100006       0         Cash loans           F            N   \n",
       "4      100007       0         Cash loans           M            N   \n",
       "\n",
       "  FLAG_OWN_REALTY  CNT_CHILDREN  AMT_INCOME_TOTAL  AMT_CREDIT  AMT_ANNUITY  \\\n",
       "0               Y             0          202500.0    406597.5      24700.5   \n",
       "1               N             0          270000.0   1293502.5      35698.5   \n",
       "2               Y             0           67500.0    135000.0       6750.0   \n",
       "3               Y             0          135000.0    312682.5      29686.5   \n",
       "4               Y             0          121500.0    513000.0      21865.5   \n",
       "\n",
       "            ...            bureau_AMT_CREDIT_SUM_OVERDUE_min  \\\n",
       "0           ...                                          0.0   \n",
       "1           ...                                          0.0   \n",
       "2           ...                                          0.0   \n",
       "3           ...                                          NaN   \n",
       "4           ...                                          0.0   \n",
       "\n",
       "  bureau_AMT_CREDIT_SUM_OVERDUE_sum bureau_DAYS_CREDIT_UPDATE_mean  \\\n",
       "0                               0.0                       -499.875   \n",
       "1                               0.0                       -816.000   \n",
       "2                               0.0                       -532.000   \n",
       "3                               NaN                            NaN   \n",
       "4                               0.0                       -783.000   \n",
       "\n",
       "  bureau_DAYS_CREDIT_UPDATE_max bureau_DAYS_CREDIT_UPDATE_min  \\\n",
       "0                          -7.0                       -1185.0   \n",
       "1                         -43.0                       -2131.0   \n",
       "2                        -382.0                        -682.0   \n",
       "3                           NaN                           NaN   \n",
       "4                        -783.0                        -783.0   \n",
       "\n",
       "  bureau_DAYS_CREDIT_UPDATE_sum  bureau_AMT_ANNUITY_mean  \\\n",
       "0                       -3999.0                      0.0   \n",
       "1                       -3264.0                      NaN   \n",
       "2                       -1064.0                      NaN   \n",
       "3                           NaN                      NaN   \n",
       "4                        -783.0                      NaN   \n",
       "\n",
       "   bureau_AMT_ANNUITY_max  bureau_AMT_ANNUITY_min  bureau_AMT_ANNUITY_sum  \n",
       "0                     0.0                     0.0                     0.0  \n",
       "1                     NaN                     NaN                     0.0  \n",
       "2                     NaN                     NaN                     0.0  \n",
       "3                     NaN                     NaN                     NaN  \n",
       "4                     NaN                     NaN                     0.0  \n",
       "\n",
       "[5 rows x 171 columns]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Merge with the training data\n",
    "train = train.merge(bureau_agg, on = 'SK_ID_CURR', how = 'left')\n",
    "train.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "290cf9aaa44d462f0f987ef35bb22b0010bd01a0"
   },
   "source": [
    "### Correlations of Aggregated Values with Target\n",
    "\n",
    "We can calculate the correlation of all new values with the target. Again, we can use these as an approximation of the variables which may be important for modeling. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "_uuid": "8cf94533bb8df06623f70d0cb01c24156f8af689",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# List of new correlations\n",
    "new_corrs = []\n",
    "\n",
    "# Iterate through the columns \n",
    "for col in columns:\n",
    "    # Calculate correlation with the target\n",
    "    corr = train['TARGET'].corr(train[col])\n",
    "    \n",
    "    # Append the list as a tuple\n",
    "\n",
    "    new_corrs.append((col, corr))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "774fe7e4c3e2708de823d3f94f65137b56bc2c2a"
   },
   "source": [
    "In the code below, we sort the correlations by the magnitude (absolute value) using the `sorted` Python function. We also make use of an anonymous `lambda` function, another important Python operation that is good to know. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "_uuid": "2802a2568214396e919837353683c63867d38a17",
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('bureau_DAYS_CREDIT_mean', 0.08972896721998122),\n",
       " ('bureau_DAYS_CREDIT_min', 0.0752482510301036),\n",
       " ('bureau_DAYS_CREDIT_UPDATE_mean', 0.06892735266968668),\n",
       " ('bureau_DAYS_ENDDATE_FACT_min', 0.05588737984392113),\n",
       " ('bureau_DAYS_CREDIT_ENDDATE_sum', 0.05373489560102051),\n",
       " ('bureau_DAYS_ENDDATE_FACT_mean', 0.05319962585758661),\n",
       " ('bureau_DAYS_CREDIT_max', 0.049782054639973095),\n",
       " ('bureau_DAYS_ENDDATE_FACT_sum', 0.04885350261111589),\n",
       " ('bureau_DAYS_CREDIT_ENDDATE_mean', 0.04698275433483543),\n",
       " ('bureau_DAYS_CREDIT_UPDATE_min', 0.042863922470730176),\n",
       " ('bureau_DAYS_CREDIT_sum', 0.04199982481484684),\n",
       " ('bureau_DAYS_CREDIT_UPDATE_sum', 0.04140363535306004),\n",
       " ('bureau_DAYS_CREDIT_ENDDATE_max', 0.036589634696329046),\n",
       " ('bureau_DAYS_CREDIT_ENDDATE_min', 0.03428110992161599),\n",
       " ('bureau_DAYS_CREDIT_UPDATE_max', 0.028233546896535826)]"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Sort the correlations by the absolute value\n",
    "# Make sure to reverse to put the largest values at the front of list\n",
    "new_corrs = sorted(new_corrs, key = lambda x: abs(x[1]), reverse = True)\n",
    "new_corrs[:15]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "581f41a665ed3b23d8b28a6e3050a11bc8a68dcf"
   },
   "source": [
    "None of the new variables have a significant correlation with the TARGET. We can look at the KDE plot of the highest correlated variable, `bureau_DAYS_CREDIT_mean`, with the target in  in terms of absolute magnitude correlation. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "_uuid": "be155ee8fd05290d11a205b95f4b553b8ade06e3",
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The correlation between bureau_DAYS_CREDIT_mean and the TARGET is 0.0897\n",
      "Median value for loan that was not repaid = -835.3333\n",
      "Median value for loan that was repaid =     -1067.0000\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAzUAAAGECAYAAAAP5Q4VAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzs3Xd4VGX2wPHvmZqENJLQBQVBQVABQaSJimsBF1HXgn3V1bX/XNeCDcUCuGvZteyuYl/XBiiiKAoqimABBaUJCCIlSAkJqVPf3x93EnInhQBJZiY5n+fJQ+57y5w7dybcc98mxhiUUkoppZRSKlE5Yh2AUkoppZRSSu0PTWqUUkoppZRSCU2TGqWUUkoppVRC06RGKaWUUkopldA0qVFKKaWUUkolNE1qlFJKKaWUUglNkxqlVNwQkc9EZHKs41CquRCRjSJyewMd+78i8mFNyw3wevNE5N8NdXylVHzTpEYppRqZiLwoIibyExSRPBFZICLjRCSrhn28IrJdREpFJKdSuYjIxyLyuYg4ovbpLyIBEflDZPkIEXlbRHJFpCxyQztDRI7ci9hzROTvIvJT5Bi/ichcEblQRFyRbf4bdX4bIufcPupYlber/JNfwzZBEdkRea/uFpGWUcd7QERWRn6fV8OxK/8cUNfzThQi4oo6x1IR+VVEpovIWdXs0gd4oo7HPm4v37drgTF1jb2uROReEVlTzapRwK31/XpKqcSgSY1SqskTEbeISKzjiPIF0A7oBAwFnsW6AVwmIodUs/1ZwHrgU+DS8kJjzaB8KdCTSjd0IpIC/Bf4rzFmioi0iezrw7r5OxQ4F/gesCUHNRGRTpHtTwfuxbohHgK8CNwOdK+0+aeR8zsQuBA4Gni9msOWb1f5J/r8y7fpBBwLTI4cc6mIdK0h3FFRxwS4Lqps857OOYH9GescDwXOA1YAr4rIa5WTX2PMNmNMcX2+sIi4I8cuMMbsrM9j18YYk2eM2dVYr6eUii+a1Cil4o1DRCZGaiV2ichkEUkuX1ldEzURuUtEfqm0/KKIzBaR6yPlPqBFZN31IrIyUsuwWkTuLK9hiKw/X0S+FpGCSAzvV04yROSgyNPqIVExrBGRe/fiPP3GmC3GmM3GmGXGmOexbvxLgOqa0FyFlTy8CPyp8gpjzCbgSuA+EekTKf474ASujywPxUpe/miM+dYYs94Y86Ux5h5jzGd1jPk/kWMeZYx5zRizwhiz2hjzAtAXWFfN+W0yxszFStqGiEiLGt6Hyj9ba9im/L16DhgABICnqws0coNbccxIcUHU64RrO1kRuSLyORkuIksjtR6fiki7SK3FYhEpitSUtYva9+RIjVKpiGwSkeekUi1cpBZtlohsE5FCEflGRE6KOsZGEblHRJ4QkZ2RWrGHJapGrgbl5/qrMWa+MeZ2rOSm/Kfya9xeafnMyHmViEi+iHwlVg1fV6zkEmBD5DswO7LPf0XkQxH5PxFZD/hEJElqaG4mIreIyObIa7wV9b5U2UdELhWRYPk1AcYBB1eqjborss7W/ExEPJH3a7OI+CPX8NxK68trtf4sIq9GruUGEbm5Du+vUirOaFKjlIo3fwCysW7CL8B64j5pH45zNHACMBo4EiiLJB1/BcYCPYAbsZKFcZX28wL3Y92k/w4IAe+LiGcfYtgrkafM/wKOE5FW5eUi0h04BngNmA60EpHjovadCrwK/FdEzgCuAC4wxhRFNskFBDi3jjfFNpF4Tgb+Wd3TcGOMv6Yn/iLSAaumKRT52W/GmHys5O9EqaHJXj1xA3cBl2F9JjsBb2B9Zq7Cqjk6CCuJBCCSnLyNVVN2OHAG0A2YUum4aVjXaxhwFDAHmFFNzdNNwK9Yn+ebgL9gfS/2mjHmHWAlcE516yPX6Q3gZayav4FYTdNCWAlrefO1vli1QGdX2n0QVq3dKKzvm7+GMAZFfk4GRkaOtTf96F4FHgF+YXeN22M1bDsJ67rdAPSKnNtr0d8drGv5CdA7cqy/i8jQvYhJKRUHXHveRCmlGlUe8GdjTAhYEXkK+4SIjN3LZjJh4KLym3qxmmPdCpxpjCl/Erwucvx/AncDRGodKojIpcAOoD/w5b6fVp0txUo+OgPbImVXATONMdsjMb2OVTPzWdS+NwCLsW6e7zXGfF2+whjzpYg8hFVj8g8RWQTMBV43xvxUh7i6ReJaXsfzOFFEirBqdpIiZZOMMWU1bFfZx8aYM+rwGuXv1UFYn5uG4ACuN8YsBRCrlvAhoLcxZkmk7Bmg8tP9e4BHjTFPlRdEPkc/i0gvY8xSY8wnUa8zVkRGYSUOlZP4T40xf4v8vlpELgNOAl7Zx/NZRtXmfeXaY90XvGmM2RgpW1HpHMrf422Var/KBYGLjTEllbavKYaLjTGFkW2ux3po0NkYs66mHcoZY0ojn5dQNTFUEJE0rD491xljypPJ+0XkaOAO7N+d1yK1fwCPisg1WO/xF3uKRykVP7SmRikVb76JJDTlvgQ8wMF7eZwVlWopwHrynAxMjTQzKYrcHP0HyCivGRGR3mJ1pl8nIoVYT8nB6hvSGMrvBE0kniTgYuClStu8CJwpItmVd4yc7yNYN5gPRR/YGHMn0Abr6fU3WE/sf6zcJKeucdXBfKwn3/0jsXyJ1Q+npu0q/1xTx9fY25j2RRB7Ircl8npLo8paV1ruB/w16nP2Q2RdNwARaS0i/xJrwIX8yDbdqfo5Wxy1vAnrGu4roeb363tgNtbDhGkicoPUfVCAZZUTmlosLU9oIsofFPSo4+vUVTesWrbPo8rnYv0tqKy+32OlVAxoTY1SKt5FP+4NV1Pmrma/6Fqd8oc4ZwOrqtk+L1Kb8xEwD+vGv/xJ8DKsxKr89auLq7oY9kUvrJvOtZHlPwBZwJSoJ99O4BLg0aj9A1jjB1TbzMsYkwdMxUru7gA+Bh7AappTm1WRuHoCM+pwHiXGmPIRqu4UkSlYNWJX1rLd3uqFdT32+IR/P4Si+t4YIBz1/hrsnwcHVhPG16o5Xvln6hWgLXALVvylWDVs0c0co5txGfbvgWQvrM9zFcaYYKTp3NHAiVhJ7yQROdMY88Eejltfgw3U9ftdV9EJXHVJXX2/x0qpGNAvrVIq3vQXEWel5YFYNx0/R5a3YjWTqaxvHY67DCgDuhhj1lTzE8J6WtwKuNMY86kxZgVW5/rKN1nlTcIqYhCR1kCHOp5fjUQkHbgamGOM2REpLh8gILo242GiBgzYW5Gb9VXYaxlq2nYbVsJ3Q6RpT3TsHqk6CEBl44DLRaT3vsYb9XqZWO/NR5H+NfFkEdCzhs9Zee3hscCTxph3jTE/Ar9hNaNrMJG+VodQSwJrLF8bYx40xgzBqkm5NLK6/ObfWe3OddNTRFIrLQ+K/Lsy8m9dvt/+OsSwGivBHxZVfiw1JHVKqcSmNTVKqXiTDTwlIv8AumA98X62Un+a2cC/ROQc4DusmoyhQK03tsaYokifkociNR4fY/0NPBzoY4y5DWvIZB9wvYg8gnWTOZFKT3Yjbfq/BG4Va04UF/BgZL+94RGRtlgJU0usgQBuxRqo4GoAETkMq/P1HeV9OsqJyH8iMRxrjIluYlOFiIzGeq/eYHety/FYtT1v1jHmP2Pd5C4SkXFYzXYCWInnrVhDUi+tbkdjzLLIqFYTgFMrrSp/H6L9ZowxUdsIVq1V+XvlxOo3EW/uBj4Qkb9jDRZQhNUc6hzgKmOMH/gJuFBEFmDVRDxA/T5ozIi8Z26gI1YH/v/D6mhf7fWOdI4/Fuu7sQVrOOheWINXgPX9MMDISM1b2T4MoSzAS5HPTw7WQATTjTHlNZOzgZtF5GqsJPpEdg9QUG4d0D7SP2YtUGyMKa28gTGmUESexPq+7wB+xHr/R2J97pVSTYzW1Cil4s0UoBCrCdjrwEzsE+q9BDwFPAksxLph+2ddDmyMuR9rBKkrgCWR17gJayQlIh3xL8Qa9WwZ1ohWf2V3k7Nyl2HdqM6PxPgM1uhie2NoZJ8NWInClcD/gF6VmmNdhTWXyrxqzmUt1vlHN+eqyY9YTYQexqpJWIg1b8v9WMnKHhljfsGam+Y9YDxWUjMfuBwr+VtR486Wh4FTRKTy0/Pjsd6H6J+MarbZgNV5+woiI4tVuhmOG8aY2Vg3432xrt0SrL5O+ewe/e0SrAT2W2Aa8C5Wkl5f/o31nq3GSmR7YI2Gd2GlZDFaPlYSPSOy37NY37eHIue1Cbgz8pMbiXtvzcfqzzUb67u9BOvzQ+Q1PsSq1bsrsm4oVsJX2TSs0eU+xKo5rWkI5tuB57ESp6VYSff5kSHGlVJNjNT8t00ppZRSSiml4p/W1CillFJKKaUSmiY1SilVj0TkgspD+Vbz0ynWMVZHRC7ZQ9zRnbcTmog493C+t+75KEoppeKFNj9TSql6FBkZrLY5Ln4xxgQbK566ioy8VtsoaOtqGiY6UYlI11pW7zDG7Gy0YJRSSu0XTWqUUkoppZRSCa1ZDulcUFCgmZxSSimllFIJKCMjI3qSXu1To5RSSimllEpsmtQopZRSSimlEpomNQqA1atXxzoEVYlej/ii1yP+6DWJL3o94otej/ij16ThaVKjlFJKKaWUSmia1CillFJKKaUSWrMc/UwppZRSSjVfxhiKiooIh8ON8npJSUkUFBQ0yms1BQ6Hg9TUVESqDHJWI01qlFJKKaVUs1JUVITX68Xj8TTK63m9XpKSkhrltZoCv99PUVERaWlpdd5Hm58ppZRSSqlmJRwON1pCo/aex+PZ61o0TWqUUkoppZRSCU2TGqWUUkoppRpRXl4eQ4YMYciQIRxyyCH06NGjYtnv9zNjxgwyMzNZtWpVxT7r16+nbdu2DBkyhAEDBnDVVVcRCAQq1i9atIiRI0fSt29fjj32WM455xyWLVsGwIQJE2yvMWTIEKZNm1bxe4cOHejXrx9DhgzhqquuavDzf/TRR+nTpw/9+vVjzpw59XJM7VOjlFJKKaVUI8rKymLevHmAlXCkpqZy/fXXV6yfOnUqAwcOZOrUqYwdO7aivHPnzsybN49QKMTo0aN5++23Oeecc9i6dSuXXnopkydPZsCAAQAsWLCAdevW0bNnTwCuueYa22sAnHnmmQCMHDmSBx54gD59+jToeQOsXLmSqVOn8tVXX5Gbm8vo0aNZtGgRTqdzv47baDU1InKKiPwkImtE5PZq1ntF5I3I+q9F5KBK68ZGyn8SkZP3dEwR+UJEFkd+NovIOw19fkoppZRSSu2voqIivv76a5544gmmTZtW7TZOp5OjjjqK3NxcAJ555hnGjBlTkdAADBw4kNNOO61RYt4bM2fO5KyzzsLr9XLQQQfRpUsXFi1atN/HbZSaGhFxAk8BvwM2At+KyLvGmOWVNrsc2GmM6Soi5wGTgHNF5DDgPKAn0B6YLSKHRPap9pjGmKGVXnsqML2BT1EppZRSSiWozBc21evx8v/YYZ/3ff/99xk+fDhdu3alZcuWLF68mN69e9u2KSsrY+HChUycOBGwaj/GjBlT63Gffvpp3njjDQAyMzN577336hRPYWEhp556arXrJk+eTPfu3Rk7dixffPFFlfVnnXUWN910k60sNzeXfv36VSy3b9++IjnbH43V/OxoYI0xZi2AiLwOnA5UTmpOB+6N/D4FeFKswalPB143xviAdSKyJnI89nRMEUkDTgD+2EDnpZRSSsWVUNiwK2DY6QtX/JQEDckuIaXST6rbQZtkBy5H3eeBUEo1vKlTp3L11VcDVvOwqVOnViQ169atY8iQIaxdu5ZRo0bRq1evao8xfPhwCgsLOf7445k0aRJQffOzukhLS6toKleTCRMm1Pl4xpi9jqEuGiup6QBsqLS8ERhQ0zbGmKCIFADZkfKvovYtT3/3dMwzgDnGmF37Fb1SSikVh4wxrC4I8tVWP/O3+Phqq5/1hSHqesvgEujQwkmnVCed0lwcmOrk8Cw3/Vt7yEnav/btTU4wiOzchuzYiiN/O/h9EAggAT8E/EjAj3E6wZuMSUqGJOtfk5yKaZmDaZkDbh1CWNUuLy+Pzz//nOXLlyMihMNhRITx48cDu/vUbNmyhdNOO42ZM2cyYsQIunfvzpIlSxg5ciQAc+bMYfr06Xz44Yf7HVN919S0b9+eTZt214xt3ryZdu3a7XecjZXUVPcYKPpvbk3b1FReXX+g6GOOASbXFtjq1atrW92s6HsRX/R6xBe9HvGnuV4TY2BRgYNpW1x8m+8kP7jvNS1BA+uLQqwvCsEWv21dx6Qwh6eF6ZUeZkBmiE7JtadKTeZ6hMMkbd9Mi41rabFpLd4dv+Ep2IG7KB/ZzyfMgRZpBNKz8Ke3xNeyNWU57Slr1Y6ynHaEvcn1dAKWJnM9GkhSUhJer7fBjl9WVlanMoBgMEggEKCsrIwpU6Zw9tln87e//a1i/ejRo5k7dy7t27cnHA5TVlZGZmYmd9xxB4888ggnnHACF198MaeeeipDhw6lf//+ABQUFBAKhSgrK7O9RnXC4TA+n6/a9W63m9mzZ9d6ruPGjavze3HCCSdwzTXXcPnll7NlyxbWrFlDz549q2y3a9cutm7dWrHcrVu3Gl8DGi+p2Qh0rLR8ALC5hm02iogLyADy9rBvjccUkWysZmpn1BbYnt6g5mL16tX6XsQRvR7xRa9H/GmO16QsaJiyroR/LS9i2c5gg7/ehjIHG8oczNxmLR+e5eaMzsmccVAyndPttw8JfT3CYRxrV+BavADHz8txrl2BlJU2yEu5iwtxFxeSkru+ahhZrQi3O5BwhwMJtz+o4l9S0/f6dRL6ejSSgoICkpKSKpb3pw9MXZSVldlerzKXy4Xb7SYpKYnp06dz00032bYdPXo07777LjfeeCMOh6Ni3RlnnMEjjzzCd999x6BBg3jxxRcZN24cubm5tGrViqysLG677TaSkpJwuVw8++yztoEHXn31VQ488EAAHA4HXq+3xhjrU+/evTnzzDMZNmwYLpeLRx55hBYtWlTZLj09nY4dO1ZzhOpJQ7Vrs72IlaSsAoYDm4BvgfONMcsqbXMtcLgx5s+RgQLONMacIyI9gf9hJSjtgTlAN6wanBqPKSJ/BgYaYy6JjqegoKDhTzrB6B/A+KLXI77o9Yg/zemaFPjDPLWsiOdXFrO9rG4zbKe7hUyvg5aRn2Sn4AsZSoKG4qChNGj1udnh27sZu8v1znZz9sEpXNA1hUyvI/GuR3ki881nuL6diyNv6573iZFwRkvCB3Qh3Kkr4Y4HW/+26wSump9LJ9z1iIGCggIyMjIa7fVqS2pU9Wq7RhkZGVWqqBulpibSR+Y6YBbgBJ43xiwTkfHAQmPMu8BzwCuRgQDysEY8I7Ldm1gDAASBa40xIYDqjlnpZc8DJjbG+SmllFL1zRjDO7+UcvvXBfxWWnPykeYWjm7t4ZjWHga29dI3x02Kq24zNhQHwmwoDvFrYYhfi4KszA/y7TY/S/MChGp5/Ld4R4DFOwqY8N0uzu+WwikpQiLcQkveNtyzp+FaMGefEplwRktMdltMVitMUgq4PRi3x+or43ZDKAS+Uqump6wU8ZUiRQXIjm1IwY59ar7mKNiJo2ARLNs95K1xuQl37EKoSw/CXboT6tID07YjOHROddV8NUpNTbzRmpqq9KlOfNHrEV/0esSfpn5N1hcGueWrfD7a6Kt2vUNgZKckruyRysA2nnofwaw4EGbxjgDfbvUzZ1MZX/7mJ1zL/5yC4ZSOyVzbK5XBbTxYg5fGD8fGdbg/eAPXgtlIaM9N90xaBqGDexLqehjhgw4h3Ko9JqsVePajD0YwiORvR/K24di+BUfurzg2r8ex6Rfkt41IeN9qzSpiTmlBqHMPQj37si69DQcMPl6TnFpoTU38i8uaGqWUUkrtWTBseHpZERO+L6S0mqqSdLdw0SEt+FOPFhyU1nD/hbdwOxjc1svgtl7+74g0fisJMWN9KW//Usr8Lf4qo/IYhA82lPHBhjIGt/VwR590BrdtuE7YdWIMjp9+wDPzNVxLvqp9U08SwSOPIdRnEKFuvTCt2kF9J2YuFyanLSanLeFDDrevCwZwbNmI5K7HsWk9js2/WP9u2YAEA3U6vJQU41q2ENeyhXQHwm88QahXP+vniAGY9Jb1ez5KxRlNapRSSqk4sKk4xGWf5fH1Vn+VdW4H3NgrjRuPSCXN3fhP39ukOLmiRypX9EhlS0mI/64uYfKKIrZU0yzuyy1+Rn6wnWHtvNzRJ40BbRo/uZHN6/H+7ylcP35T4zbliUzw6OMIHTkA6nn0sb3ichM+oDMc0JlQ/0rloSDy2yacG37GsX4Njg0/4/h1DY78HXs8pKMwH8eC2bgXzMY4HIQOP5rg4JMJ9hm0fzVOSsUpTWqUUkqpGPt0UxlXzN1Zbcf9gW08PDYok+6Z7hhEVlXbFCd/PTKNG3qlMm1dKU8vK+KHvKq1CXNzfczN9TG8g5exfdLp16oR5mgpKcIz/WXcH09FQqFqNwm3bo//1HMJDj4ptolMXThdmPYHEmx/IAw4oaJYCvJwrFuJ8+cVONautEZsKymq8TASDuNa8hWuJV9hklsQPPo4AoNPtmqM4qypoFL7SpMapZRSKkbCxvD3JYVM+L6wSpOuTI8wvn8GF3ZLwRGHN54ep3Be1xTOPTiZNxet5Y0dGXyyuWofoDmbfMzZtI2TD7CSm945DZDchMO4vvgAz5TJOHbtrHaTUJce+EecR+ioIeBI7IlFTUYWod6DCPUeZBWEw8iWDThXLMa19Ftk2SKcvuqHpZbSYtxz38c9933rPfn9hYT6DNLkRiU8TWqUUkqpGMgrC3Hl5zuZvalqInBiBy//GtqSVsnxf/MtIvTNCHNuvxwW/OZjwveFfJ5b9ZxmbfQxa+M2RnRK4vbeaRyRXT/JjWxeT9LkiTh/XlHt+tChR+I78zLChx7RdG/cHY6KGp3g8NNZ/dNKDnUEcf34Lc6Fc3Fu+qXa3ZxrV5D8jzsJdTyYwO8vINh/WMInfIkiLy+PUaNGAbB161acTifZ2dkAfPLJJ8yaNYuLLrqIb775hkMOOQSA9evXM2DAALp27UogEKB37948+eSTuN1WLe6iRYu45557yM3NJTU1lbZt2zJu3Dh69uzJhAkTePnllyteA+Avf/kLjz76KADr1q2jXbt2JCUl0bNnT/7zn/806LlffPHFfP/995x//vm2iUb3hyY1SimlVCP7KT/AOR/vYH2RvYmUAHf0SePmI9PisnZmTwa28fLuKV6+yPXx0Pe7WPBb1f5BM38tY+avZZzZOZk7+6RzcMY+3oqEw7g/mopnyrNIoOrrhLNa4z/vaoJHH9d0k5maOJyEu3XH360XnHEpjl/X4PryI1wLZldbk+Xc8DPOp8cTbtsR/5l/JHj08c3vPWtkWVlZzJs3D4AJEyaQmprK9ddfX7F+6tSpDBw4kKlTpzJ27NiK8s6dOzNv3jxCoRCjR4/m7bff5pxzzmHr1q1ceumlTJ48mQEDBgCwYMEC1q1bR8+ePQG45pprbK8BcOaZZwIwcuRIHnjgAfr06dOg5w3g9Xq58847WbFiBStWVP8wYl9oUqOUUko1os9zfVz0yQ4K/PYGZ9leB88d15Lj2if+sK9D23mZ2TaHz3N9PPR9YbWDH0xbV8r0X0q5+JAUbu2dTruUutcQyNbNJE2ehPOnJVXWGbeHwMgx+EeMAW/iv5f7TYTwgd3wH9gN/7lX4Vy6EPcn03EtXlBlU8eWDSQ9PZ7g3Jn4Lvk/TJsDYhCwKioq4uuvv2bGjBmMGTPGltSUczqdHHXUUeTm5gLwzDPPMGbMmIqEBmDgwIGNFvPeaNGiBQMHDmTt2rX1elxNapRSSqlG8urqYm78Mp9gVAeao1t5eOH4LDq0aDpNf0SEYe2TOLadl082+3jou10s2m4fUCBk4IWfSnh9TSl/PqwFNx6eRqa3ltHdjMH16bt4X/8X4iursjrYZzC+C6/H5LSt79NpGpwuQkceQ+jIY3D8ugb3jFdxfftZlUlBXcsW4rzzj/hPu5DAyDHW5KJNXOolx9Xr8Ype+myf933//fcZPnw4Xbt2pWXLlixevJjevXvbtikrK2PhwoVMnGjNM79y5UrGjBlT63Gffvpp3njjDQAyMzN577336hRPYWEhp556arXrJk+eTPfu3Rk7dixffPFFlfVnnXUWN910U51eZ39pUqOUUko1MGMMD35fyN+XFFZZd2G3FB4dmInH2TSb+4gIwzskcUJ7Lx9t9PHAd7v4MWq0tNKQ4bEfi3h5VQl39U3n4kNScEZPKLorn6TJE6udc8aktMB34Y0EB/1Om03VUbhTV3zXjsOf+0c87/0P14KPbSPGSSCA9+0XcC+Yje+Smwgd1jeG0TYvU6dO5eqrrwas5mFTp06tSGrWrVvHkCFDWLt2LaNGjaJXr17VHmP48OEUFhZy/PHHM2nSJKD65md1kZaWVtFUriYTJkzY6+PWN01qlFJKqQbkDxmunbeTt9ZWHY3qnqPSuenwVKQZ3IiLCCd3TOJ3B3h5e10pD3y3i3WF9j5FO3xhblqQz7Mri5hwdCbD2lvzqTh//BbvsxNwFORVOW7w8P74LrsFk9W6Uc6jqTHtOuH70+34Tzsf78uP41r+nW29Y8sGkif9Bf/IMfjPuhyceuvYkPLy8vj8889Zvnw5IkI4HEZEGD9+PLC7T82WLVs47bTTmDlzJiNGjKB79+4sWbKEkSNHAjBnzhymT5/Ohx9+uN8xaU2NUkop1cwVB8Jc/Gkec6JGOPM64d9DW3JG55QYRRY7DhHO6pLCqIOSeWVVCZMW7+K3qEk8l+8Mcvqs7Yzq4OSZ3KnkfPJWleOYpGR8Y64lOGyk1s7UA9OuE2W3PoJr/sd4XnsaR2G+bb3n/ddwrPuJsqvvgfTMGEXZ9E2fPp3zzjuPxx9/vKJsxIgRLFiwgA4dOlSUlY9s9thjjzFixAj+9Kc/MXz4cIYPH17Rr6akpKReYtKaGqWUUqoZ2+kLc87H2/l2m72pVbbXwWsnZnF06+Y9q7vbIVzWvQXnHpzMk8uK+MePRZRU6mx0aPFm7pr+FDnHzvuGAAAgAElEQVRFv1TZN9S1J2V/vgvTql0jRtwMiBAcfBLB3gPxvvUM7k9n2Fa7ln9Hyrg/UXbdeMIH94hRkA1jf/rA1KcpU6ZUqdkYNWoUU6ZM4cYbb7SVn3baaUycOJH58+czaNAgXnjhBcaNG0dubi6tWrUiKyuL2267rWL7yn1qAF599VUOPPDAhj2hGhx++OEUFhYSCAR4//33mTZtGt27d9+vY4ox0dN9NX0FBQXN76T3YPXq1XTr1i3WYagIvR7xRa9H/In3a7K5OMRZH21nRX7QVt4lzcm0k3M4KK1pPVOsj+uxqTjEfYsKeHNNCZflfsZja16hRdhew2XEQWDURfhPv0ibQdWivr4fjp9+IOnp+3Dk77CVG5cb34U3EDzutIStJSsoKCAjI6PRXq+srIykJB2Nb2/Udo0yMjKqfPBqGWJEKaWUUntrTUGAk2duq5LQHJHl5sORrZpcQlNfOrRw8kxfF7l5/+aZVZOrJDS/erM5/sg7ubHdGewK6e1LYwgfegSl9z1D6NAjbeUSDJD04iN4XvkHhMM17K1U49K/CkoppVQ9Wb4zwKkzt7MhalLNwW09zDg1h9bJTWfI5vrmWLmYlLsvp9WPVdvuv9VqAH37TWBeZneeXVnMoHe28nmur5qjqPpmMrMpvfUR/KecU2WdZ847eF98VBMbFRf0cZFSSilVD5blBTh91na2l9lv8EZ2SuK5YVkkuRKzmU6DCwbwvPMS7vderTJfStibxLNHX861joG2Zk4bi0OM+nA71/Rswd19M0jW97ZhuVz4x1xDuEsPvM9Nss0R5J77HmDwXXozOPRZuYod/fQppZRS+2lpXoBRH1ZNaC7qlsJLx2tCUxPHL6tIHncVnhn/rZLQhA46hNLxz3LBFWfz9sk5HJRWtZbr6WXFHPfuVr7f7m+skJu14IDjKb3nacIZWbZy99z38b7wd62xUTGlSY1SSim1H36MJDQ7fPYbuit7tOCfgzNxRU8iqazamWkvkDz+apwb11ZZ7R8xhtK7n8K07QjA8R2SWDC6Ddf2TCX63fypIMiJ723j4cW7CIV1HKCGFj6gC6W3P1Y1sfl8Jt7nHoZwqIY944vD4cDv12Q4Xvn9fhx7WfOnzc+UUkqpffTDDj+nz9rOTp/9ZvrPh7VgwtEZzWJSzb3lWL8a7+SJOH/9ucq6cGY2vivHEurZr8q6ZJfw4NEZnNopiau/2GnrtxQy8ND3hXy5xc+zw1pq36UGZtofSOnYx0meeJNtZDT3vA8Bg+/y2+K+KVpqaipFRUWUlladFLch7Nq1i/T09EZ5rabA4XCQmpq6V/toUqOUUkrtg5/yA9UmNFcf1oKHNKGpyleG591XcH/wOhKq+jQ/MPBEfBfeAKm13/gNaevly9Nbc+c3Bbyy2j654NxcH0Onb+XZYVkc2655zwPU0Ey7TlZiM+EmHPnbK8rd82YRzmlH4IxLYxdcHYgIaWlpjfZ6W7dupWPHjo32es1RfKfRSimlVBzKLQlx1kc7qiQ01/TUhKY6zu/nk3LHJXjee7VKQhPOaEnpjQ/g+/Nde0xoyqV7HDwxpCWvDc8iJ8l+K/NbaZjRs7YzSZujNTjTtiOlYx8n3DLHVu5950Wci+fHKCrVXGlSo5RSSu2FXf4wZ3+8g43F9pvz63qm8mB/TWgqkx2/kfSPO0l+/A4c23+rsj4w8ERKHnqRUN8h+3T8Uzsl88XprRnc1mMrDxuY8H0hZ328g50+7bzekEzbA6zEJi3TVp707weRLRtiFJVqjjSpUUopperIHzJc/GkeS/MCtvJLDknh/v7pmtCUCwZxv/8aKbdfguu7L6usDmfmUHrD/ZHamf2b1b1dipPpJ+dwy5FpVQYR+Gyzj+EztrK6IFDtvqp+mDYH4Lt2HKZSPxopLSbpH3dDaUkteypVfzSpUUopperAGMP1X+7ks832SR9P7pjEIwMzNaGJcKxcQvLdV+B98z+Iv8y2zjgc+E8+m5KJLxM6ami9vabLIdzZN51pJ2VXaY62tjDE8Pe28cmmshr2VvUh1KMP/vOutpU5N/9C0uSJYLQZoGp4mtQopZRSdXD/d7t442f7SElH5bh5flhLHbYZYFc+3mcnkDLhRpybf6myOnTwYZTe9wz+86+F5JQGCeH4DknMO701x7S2N0fb5Tec/fEOnllehNEb7AYTOOkPBAaeaCtzLfwc93v/i1FEqjnRpEYppZTagzd+LuHRH4psZZ3TnLx+YjYt3M38v9JwmOzvPqfF7RfhnjerymrTIo2yP/6V0rueJNypa4OH0zbFyfRTcji/qz1xChm49esC/vpVgQ4g0FBE8P3xr4SirrNn6mScP34bo6BUc9HM/xIrpZRStVuZH+Cm+fm2spwkB1NPyqFVM58PRTavJ3nCjXSa+QpSXFhlfWDIKRRPfIXgcac16rwlXqfw1JBM7u+XXqWfzXMri7li7k4Cmtg0DG8SZTfcj2mxeyQ7MQbvc5OgtDiGgammTpMapZRSqgbFgTCXfppHSXD3DXCSE944MZsu6c14qreAH8/bL5By1+U4V/1YZXWow0GU3PFPfH+6HdIzqzlAwxMRrj88jddOzCLVZU9t3v6llIs+yaMsqIlNQzCt2lF2zT0Y2X2b6di5Hc9bz8YwKtXUaVKjlFJKVcMYw80L8lmZH7SVP3xMJke18tSwV9Pn+OkHUu6+As87LyEh+3tjPEn4zv0zpeMnEz70iBhFaHdKx2Q+Oq0VB7Sw16p9uKGMc2fvoDigQz43hFCvfgRGnGcr88x5B0c1SbBS9UGTGqWUUqoa/11dwutRAwOce3AyF3VrmE7ucc/vw/Pff5Ly0A04cn+tsjp45DGUTHzJupF1xVct1mEt3XwwIocuafbEZm6uj7M+2kGBXxObhuAffQnhNh1sZUnP/w0C/hhFpJoyTWqUUkqpKEvzAtzylb0fzaEZLh5tpkM3O35ZRcq4K/F8PK3KunBGS9adeSVlN03AZLeJQXR10zHVxcwRreiRaU+4vtrqZ9SH28nXSTrrn8eL749/tRU5cn/FM+PVGAWkmjJNapRSSqlKCiP9aMpCu8tSXMJLJ2Q1v5HOwiHc771K8vhrcGxeX2V1YNhplEx4mfzD+kMCJHttU5y8d2oOR2a7beVLdgQ4b/YOSoKa2NS3UI8+BI4dYStzv/cqjo3rYhSRaqqa2V9npZRSqnZ3f1PAml32viKPDMyke6a7hj2aJtmWS/KEm/C+9WyVvjPhVu0pGfs4vsv+Ci3SYhThvslOcvLuKTkMiJrL5qutfv74aZ6OitYAfOddTTijZcWyhIJ4n/8bhEO17KXU3tGkRimllIr4bHMZL64qsZVd1C2FMV2bVz8a5/fzSbnnCpyrfqiyLnDsCErun0y4e+8YRFY/MjwOpp2UzcA29sRm1kYf18/bSVgn6KxfLdLwXXSjrcj583Lcc6bHKCDVFGlSo5RSSmE1O7v+S3s/mm4ZLiYdkxGjiGIgHMIz9TmSH78DKbHPKWJS0ym9/n58l98KyYmf5LVwO3hteDY9W9r72Lz+cynjFu6KUVRNV6jfMIJ9BtvKPFOeRfJ3xCgi1dQ0WlIjIqeIyE8iskZEbq9mvVdE3ois/1pEDqq0bmyk/CcROXlPxxTLgyKySkRWiMgNDX1+SimlEtt9C3exoWh3cxgBnhycSYqrmTz/Kyog6ZHb8bz7SpVVwcP7U/LgC4T6DY1BYA0n02tNonpgqn1UtCeWFvHPH6tOJqr2gwi+i2/EJO1OiKWsFHc1nzel9kWj/KUWESfwFHAqcBgwRkQOi9rscmCnMaYr8BgwKbLvYcB5QE/gFOBpEXHu4ZiXAh2B7saYHsDrDXh6SimlEtwXuT4mr7TXTFzdswUD2nhjFFHjcqz7iZRxV+Ja+q2t3Dgc+MZcQ9nND2Mys2MUXcNqm+Jk2kk5tEqy3xLds3AXb/1cUsNeal+YrNb4z7rMVub+bAaydXOMIlJNSWM9fjoaWGOMWWuM8WMlGadHbXM68FLk9ynAcLHGzTwdeN0Y4zPGrAPWRI5X2zGvBsYbY8IAxpitDXhuSimlElhxIMz1X+60lXVJc3JX3/QYRdS4XPM/JvnB63Bs/81WHs5oSeltjxE45ZyEGNlsfxyc4eKt32WT5raf53Vf7mTRNp1TpT4Fjh9FOKdtxbKEQnimPR/DiFRT0VhJTQdgQ6XljZGyarcxxgSBAiC7ln1rO+bBwLkislBEPhCRbvV0HkoppZqY+7/bxS+FUc3OhrRs+s3OjMEz7QWS/vMgEgjYVoW69qL0vmcJdz8yRsE1vt45Hl4dno2n0mX3heCCOTvYXKyjdNUbtwf/mVG1NQtm41i/OkYBqaaisab8re4RT/TQIjVtU1N5df/blB/TC5QZY/qJyJnA80C1DYFXr9YvUTl9L+KLXo/4otcj/tTHNVmyy8F/lnup/F/NOe0CtCr8ldVNuEuFBAN0mvEiqcu+qbJua//hbD7xD5jtO2H7zmr2rl5T+I60A+442Mm9q3c3O9xSGuasmZt45nAfSc6a9403cX09cg6ke+sOJG/dVFEUeOkfrB1zYy07Jb64viYJoFu32usoGiup2YjVx6XcAUB0A8rybTaKiAvIAPL2sG9N5RuBqZHf3wZeqCmwPb1BzcXq1av1vYgjej3ii16P+FMf1yQQNlw0fSuG3XOwHJjq5NHh7Zr2JJu78kn+5104Vy+1FRuXG99lt5Ay+CS67uUhm9J35P+6wQ5vAU8sLaooW1Hk5PEt2Uwe1hJJgKZ4CXE9LrgOHhtbsZjx81IODRUn9FDhtUmIa5LgGuuv9rdANxHpLCIerI7/70Zt8y5wSeT3PwCfGGNMpPy8yOhonYFuwDd7OOY7wAmR34cBqxrovJRSSiWofy0rYmW+fVLJJ4a0bNIJjWxeT8r4a6omNGkZlN72KMHBJ8Uosvhy71HpnHSAfZCIqetKefSHohr2UHsrdOQxhA453FbmffMZ0DmC1D5qlL/ckT4y1wGzgBXAm8aYZSIyXkRGRTZ7DsgWkTXAX4DbI/suA94ElgMfAtcaY0I1HTNyrInAWSLyIzABuKIxzlMppVRi2FgUZOJie/uyC7qlcGy7pjvamWPNMlIeuA7HNntDiXC7TpTc/TThqBvM5szpECYPy+LQDHuDlvu/28V760tjFFUTI4Lv7CttRc6fl+P8/ssYBaQSXWM1P8MYMxOYGVV2T6Xfy4Cza9j3QeDBuhwzUp4PjNzPkJVSSjVRY78poCS4+4lwpke4r1/THe3MueQrkp4ch/h9tvLgYX0pu+4+aJEWo8jiV7rHwWsnZjP8va3s9O3+rFz1+U5mjXTRK8sdw+iahvAhhxPsPQjX4vkVZZ63JlPaeyA4EqgDk4oLTbeOXSmllKrGRxvKmLG+zFZ2b78MchKpF/hecH35EUmP31EloQkcO4Kymx/WhKYWXdJdvHhcNs5K3WiKg4Yxc3awrVRHRKsP/rOvwFTqp+Tc/Auu+bNjGJFKVJrUKKWUajZKg4Zbv863lR2V4+biQ1Jq2COxuT94g6RnHkLCYVu5b/Sl+C67BVyN1mAjYQ1r72XSgAxb2YaiEBd/moc/pP0/9lf4gC4EB9n7cnne/x9EfWaV2hNNapRSSjUbj/1YaJuTxiHwyMBMHAkwotVeMQbPG//G+/q/7MUilF18E4EzLm3yE2rWpyt6pHJ59xa2sgW/+bl5QT5GO7bvN/8Zl2Icu29JHZvX41y8IIYRqUSkSY1SSqlm4eeCII//YB8c4IruLeid44lRRA0kHMb70mN4Zr5uKzYuN2XXjCM4/PQYBZbYJg7IYGhb+2flldUl/Ht5cYwiajpMq3YEB5xgK/PMfC1G0ahEpUmNUkqpJs8Ywy1f5eOv1KKldbKDO/s2scEBwiG8L/wd96f2WRNMUjJlN08idPRxsYmrCXA7hJeOz+KgNHvfqzu/LeCTTWU17KXqKjDiPNuyc/VSHKt+jFE0KhFpUqOUUqrJm/5LGZ9stneUf7B/BhmeJvTfYCiI99mJuD+3DwoaTsuk9PbHCR3WN0aBNR1ZSU5eG55Nmnt3072wgUs/y2N1QSCGkSW+cKeuBA/vbyvzvK+1NarumtBfc6WUUqqqwkCYsd/YBwcY2tbDH7okxyiiBhAM4v33g7jnf2wrDmdkUTr2ccKdD41RYE1Pj5ZuJg/LonKPpF1+w5jZeeT7tHP7/giMPN+27Fo8H9n0S2yCUQlHkxqllFJN2sTvC8kt2X2z6XZYgwNIU+koHwyQ9PR9uL/51FYczmpF6R3/xHQ4KDZxNWEnd0yqMq/Rml1BLvssj2BYBw7YV6HuvQlFJeDRfcOUqokmNUoppZqspXkB/r28yFZ2Q69UDslsIhMnBoMkPXUvrkVf2IrDOW2shKbtATEKrOm7vlcq5x5sr+37ZLOPu74tiFFETYAI/pFjbEWuBbORvK0xCkglEk1qlFJKNUlhY7h5QT6VpxLpmOrk5iObyGST4RDeZyfg+u5Le3Hr9lZC06pdjAJrHkSEfwxqSf9W9gT538uLeXmVjoi2r0JHDSXcpkPFsoSCuGdNiWFEKlFoUqOUUqpJenV1CV9v9dvKHh6QQYqrCfzXZwzeFx/F/dUcW3G4XUdKx/4Dk90mRoE1L0ku4b8nZNMhxT4i2s0L8pm3xVfDXqpWDif+U8+1Fbk/mwHFhTXsoJSlCfxlV0oppezyykKMW7jLVnZqxyRO7dQEBgcwBs9rT+Oe+76tONy6PaW3PYbJahWjwJqnNilOXh2eRbJzdx+tQBjOn72Dxdv9teypahIcfDLh9JYVy1JWivuTd2vZQylNapRSSjVB4xftIq/SSFTJTmHigIwYRlR/PG+/iGfWW7aycMscSm99BNMyJ0ZRNW+9czw8PTTTVrYrYDjzox2s2KlDPe81j5fASWfZitwfT4WgvpeqZprUKKWUalIWb/fz0qoSW9ktvdM4MM0Vo4jqj3vm63imv2QrC6dlWgmN9qGJqTM6p3BX1GSueb4wo2dtZ+2uYIyiSlyBE07HJO2uWXUU5FUZEEOpyjSpUUop1WQYY7jt6wIqD6rbNd3FdT1TYxZTfXHN/xjvG/+2lZmUVMpu/Tum/YExikpVdvMRqdzYy/5Z+600zKgPt7OhSBObvdIijcDgk21F7jnTYxSMSgSa1CillGoy3lpbWmVwgIkDMvA4E3tOGsdPP+B97mFbmfEmUXrzJMKdusYoKhVNRLi3XzpXdG9hK99YHOL0D7ezpSQUo8gSU/CE023Lzp+W4NiwNkbRqHinSY1SSqkmoSgQZtxC+xwhp3RM4sQDkmIUUf2QLRtJ/uddSKX+BMbpouz/HiLctWcMI1PVEREePiaD87um2MrXFob4/Yfb2VysiU1dhQ/oTKj7kbYy9ydaW6Oqp0mNUkqpJuHRHwrJLdk9OIDHAQ8dneCDAxQVkPzo7UiRfSQ33+W3Ejqsb4yCUnviEOGJwZmccZB9tL3VBUFGfrCNX7UpWp35h59hW3bN/whKdR4gVZUmNUoppRLe2l1BnlxaZCu7tmcqXdITeHCAgJ/kf9yN47eNtmLf6EsJDj4pRkGpunI6hP8c25KTO9prCtcVhhgxczvrdPCAOgn1HUI4M7tiWcpKcX/5UQwjUvFKkxqllFIJ785vCvDvrqShXYqDm49Mi11A+8sYvM89jHPVD7biwKDfERh9SYyCUnvL4xRePj6LEZ3sic3G4hAjPtjG6gIdoniPXC6Cx51mL5ozHYypYQfVXGlSo5RSKqHN2VTGBxvKbGX39csg1Z24/8W533sV94LZtrLQoUfiu+wWkMQe9KC58TqFl47PqtIULbckzMgPtrNc57HZo8Cw0zCO3d9n5+ZfcPy0JIYRqXiUuH/xlVJKNXvBsOGOr+2DAwxo7eHsLsk17BH/nMsW4pn6vK0s3LYjpTfcD25PjKJS+8PtEJ4d1pJzD7Z/LreWhjntg+0s2eGvYU8FYLJaEeo7xFbmnv1OjKJR8UqTGqWUUgnrv6tL+Klgd98EASYNyEAStDZD8raS9K/7EbO7LZ1pkU7pXyZAanote6p453II/xrakosPsY+Kluez5rFZtE0Tm9oETowaMOC7L5Cd22MUjYpHmtQopZRKSEWBMBO+t48Kdn63FHrnJGhtRjBA0pP3IoW7a56MCGVX341pc0AMA1P1xSHC44My+VMP+zw2BX7D6FnbWfCbL0aRxb9Q996EK00yK6EQrrnvxzAiFW80qVFKKZWQnlxaxG+lu2s0kp3CHX0StzbD8/q/cP683FbmP+OPhA7vH6OIVENwiPDwgAyu75VqKy8MGM76aAdzN2tiUy0RAlGTcbo/nQFBHUVOWTSpUUoplXC2++GJqCGcr+nZgg4tnDGKaP+4FszB8/E0W1nwiAEEfn9hjCJSDUlEGN8vnb9GjdBXEjScO3s7szeW1bBn8xYYfBLGu3skOUf+dpyLF8QwIhVPNKlRSimVcCb/6qY4uHtI12yvgxsPT8whnB0b1+F9/m+2snBOW8quuhMc+t90UyUi3NU3nTv72D+3ZSG44JMdfLpJE5sqUlIJDvqdrcg978MYBaPijf61VEoplVBW5Qd4Z4t9Us3beqeR7knA/9J8ZSQ9dS/i330Da9xuyq67TwcGaCZu6Z3O/f3t19oXgvPn5DFvizZFixY4dqRt2blkAVKQF6NoVDxJwP8BlFJKNWf3LdpFiN2jm3VJc3LpoS1q2SN+ed74N47N621lvgtvJNz50BhFpGLh+l5pPDwgw1ZWGjKc+/EOvtbBA2zCnQ8l1P6gimUJh3FFzemkmidNapRSSiWMBb/5eP9Xe7Occf0y8DgTbwhn55Kv8Myxz7URGHIywWEja9hDNWVXHpbKhKPtiU1x0HD2xzv4Tod73k2E4NBTbEWuLz4AY2rYQTUXmtQopZRKCMYYxn1rH8K5Xys3ow5MqmGP+CW7duKdPMlWFm7dHt9FN0KCzrGj9t/VPVO5r5+9KdqugOGMj7bzg07QWSE46HeYSv3NnBvX4Vi/OoYRqXigSY1SSqmEMHuTj2+inliP75eAE20ag/e5v+HYtXN3kcNB2Z/vgqSUWnZUzcGNh6dxR9TgAQV+wxmzdrC6IBCjqOKLycwmdMQAW5nriw9iFI2KF5rUKKWUinvGGCYtttfSnNIxiUFtvTGKaN+5PpuBa/F8W5n/9EsIH3xYjCJS8eaWI9O4+Qj7PDY7fGHOmLWDTcWhGEUVXwJD7E3Q3AvmQEBrs5ozTWqUUkrFvTmbfCzcZn9KfXvvxBvCWXJ/xfu/p2xloYMPI/D7C2IUkYpH5cM9X9fTnthsLA5x1kfbySvTxCbUeyCmxe6melK8C2fUwwLVvGhSo5RSKq5VV0szNCtI7xxPjCLaR8EgSf95EPHvHs3KJCVb89E4XbXsqJojEeH+/ulcfIi9SeLK/CDnzt5BcSAco8jihNtDYNCJ9qIvdM6a5kyTGqWUUnHt080+vo2qpflTp8TrW+B+71Wc636ylfkuuAHTpkOMIlLxTkR4dGAmp3WyD4bx7bYAl3yahz/UvEf8CkY1QXP++A2SvyNG0ahYa7SkRkROEZGfRGSNiNxezXqviLwRWf+1iBxUad3YSPlPInLyno4pIi+KyDoRWRz56d3Q56eUUqr+GWOY+H2hrezkjkn0SE2smznHhrV43n3FVhbsd2yVoWmViuZyCJOHZTGkrb1mcvYmH9fM20m4GQ9lHD6wG6EDulQsSziMa/7HMYxIxVKjJDUi4gSeAk4FDgPGiEh0j8jLgZ3GmK7AY8CkyL6HAecBPYFTgKdFxFmHY95ijOkd+VncgKenlFKqgXy2ueqIZwnXlyYUxDt5IhIKVhSF01tSdulfdPhmVSdJLuF/w7M5IsttK5+ytpTbvy7ANNfERoTg0FNtRa4vPtQ5a5qpxqqpORpYY4xZa4zxA68Dp0dtczrwUuT3KcBwscbpPB143RjjM8asA9ZEjleXYyqllEpQVl+aqFqaA7z0SbC+NO6Zb+D8ZZWtzHfx/0FaZowiUoko3eNgyknZdElz2sqfWVHM35cU1rBX0xccdCLGufs9cW7+BUdUM0/VPDRWUtMB2FBpeWOkrNptjDFBoADIrmXfPR3zQRH5QUQeE5HEG/NTKaWaubm5Pr7aaq+lua13eg1bxyfZ9Aued160lQX6H0eo/7DYBKQSWutkJ9NOzqFtsv327cHvC3l+ZXGMoootk96S0JHH2Mpc83TAgOaosYZbqa5+PbpusKZtaiqvLiErP+ZYYAvgAZ4BbgPGVxfY6tU6A205fS/ii16P+KLXo3EZA/f+6AV2P4Ed3DJEWv56Vudby3F/TcJhDnlpIhLcPahBICWVlUN+TzDeY98HcX89mpDHugtX/pBEYWj3LdLNC3bi2/kbJ+ZYwz03p+uRcfCRdPnuy4plWTCH1QNOAYezlr0aX3O6Jg2hW7duta5vrKRmI9Cx0vIBwOYattkoIi4gA8jbw77VlhtjciNlPhF5AfhrTYHt6Q1qLlavXq3vRRzR6xFf9Ho0vrmby1i8yz6K0fjBbenWymp6lgjXxD3zdbyb1tnKQpfcROfeR8UoooaTCNejKekGvNXOx+hZ2ymfssYg3LPKS48Ds+lQvKF5XY/OB2HefxkpKQLAXVJI92ARoZ79YhtXJfodaXiN1fzsW6CbiHQWEQ9Wx/93o7Z5F7gk8vsfgE+M1fPtXeC8yOhonbG+y9/UdkwRaRf5V4DRwNIGPTullFL1xhjDxKi+NCd28HJUq8TpSyNbNuCZ9rytLNh3CMEBJ8QoItXUHNPGy0vHZ+Os1J4lEIYL5+SxvLCZzdjhchM8aqi96OtPYxSMipVG+dRH+shcB8wCVgBvGmOWich4ERkV2ew5IFtE1gB/AW6P7LsMeBNYDnwIXGuMCdV0zMixXhWRH4EfgRzggRZKp10AACAASURBVMY4T6WUUvvviy1+FvyWwH1pwmGSnnsYCew+B9MiDd8lN+loZ6pendwxiaeGtLSVFQUNNy7zsio/8eZy2h/Bo4+zLbsWfgHBYPUbqyap0aYwNsbMBGZGld1T6fcy4Owa9n0QeLAux4yU66MwpZRKUBO/32VbHt7BS//WiVNL4/psBs5VP9rKfBdcj8nMjlFEqik7r2sKO3xh7vymoKIsPyic+dEOZo1sRYcW8dWvpKGEDjsKk5qOFFl/P6R4F85liwgdOSDGkanG0szqJ5VSSsWzL3J9zK9SS5M489JI3ja8bz5jKwseeQzBQb+LUUSqObi2Zyo3HZ5qK9tYHOKsj7aTV97ppqlzuQgeday96BttgtacaFKjlFIqbkxabK+lOaG9l6NbJ86o/N5X/oGU7h5a13iTtNmZahT3HJXOxYek2MpW5gc5d/YOigPhGEXVuIIDjrMtu777AgL+6jdWTY4mNUoppeLCvC0+5m1J3Foa58LPcX03z1bm/8OfMNltYhSRak5EhEcHZnJapyRb+bfbAlz8aR7+UPRMGk1PqHtvwpUmtZWSYpxLF8YwItWYNKlRSikVFyZF9aU5vr2XAW0SpJamuBDvy4/bikIH9yBw4ugYBaSaI5dDmDwsi6My7E3O5mzycc28nYRNE09snK4qE9tqE7TmQ5MapZRSMfflFh9fJHAtjffNZ3AU5FUsG6cT32W3xN3kf6rpS3IJf+/h44gst618ytpSbv+6ANPEE5sqo6B99yX4fbEJRjUqTWqUUkrF3KSoeWmOa+/lmASppXGsXIz7sxm2ssDI8wkf0CVGEanmLtUFU07KpkuaPal+ZkUxf1tSWMNeTUPo0CMIZ2RVLEtZCc4fvolhRKqxaFKjlFIqpuZv8fF5rv1JasLU0vh9JL3wiK0o3K4j/t9fGKOAlLK0TnYy7eQc2ibbb/Ue+r6Q51YWxSiqRuBwVq2t+eaT2MSiGpUmNUoppWIqupbm2HZeBiZILY3n3VdwbNlgKyv74y3gSYz4VdN2UJqLqSflkOGxj7731wUFvLOuNEZRNbwqSc33C8BXFptgVKPRpEYppVTMLPjNx9wEraVx/Poz7pmv2coCx/+e8KFHxCgiparqmeXmjROzSarUEs0Af/o8j882N80b/XDXXoRb5lQsi78M55KvYhiRagya1CillIqZ6FqaoW09DG6bALUc4RDeF/6OhHaPMhXOzMZ3zlUxDEqp6h3TxstLx2fjrFRhEwjDhXPyWLy9Cc7j4nAQPPp4W5H7a22C9v/s3Xd4W9X5B/DvuVfTO87esR1n7xA2lBUIpeyw96asUlog7BFCoUBTZlv2+NFC2KGkjLBHBpBJph1nLyfeQ/Pe8/vDwtKVLdtyrHsl+/t5njxYr3SuX1vY8qtzzns6OxY1RERkiUW7ffhqR9QszcQsi7KJj33+e1BL1hhivgtuBNIyYowgstZxA114+tBuhlhtUOKMz8pQUh20KKvEiV6Cpi5fCHjrrUmGTMGihoiILBE9S3NoHwcOTYFZGrF3FxxvP2+IBScfBm2/wyzKiKhtzh6ahgemGN842OPVcdqne7G7XosxKjXpBaOg9wgffCsCfqirfrYwI0o0FjVERGS6xaU+fBE9SzMhBWZppITzldkQEZuOpTsdvgv+YGFSRG133ZhM3DDGOKO4qUbD9M/KUO3XLcoqAYRAcOKhhpBt2QKLkiEzsKghIiLTRc/SHNLHgcP6Jv8sjW3hF7CtWGSI+c66CjJiUzJRsrt3vyycXeA2xFaWB3D+F+XwaZ3ncE5twkGG2+ryBYDeiQo3MmBRQ0REpvqx1I/Pt6fgLE1tFRyvP2kIacPGIfib31mUEFH7KELgyUO7YWp/4xsJ3+z04YbvKyBl5yhstBHjIV3h4k2pqoCyab2FGVEisaghIiJTPbys2nD7oN4OHNbHYVE2bef8vyeh1FQ23pY2O7yX/hlQ+FJKqceuCLx8ZC7262k3xN/c4MGTv3SSwzltdmhjphhDy36wKBlKNP4mJiIi0/xY6sf8qFmaGROyIISIMSI5qD9/C/uC+YaY/6QLIPsOsigjon2Xblcw55juGJplM8Tv+akan2ztHGfYBCcebLitcl9Np8WihoiITDNzSdNZmsP7JvksTW0VnK/8zRDSBhYgcMI5FiVE1HFyXSreOCYX2Y7wGwsSwOVfl2NtZcC6xDqINu4AyIg3TdTNRRDlpRZmRInCooaIiEzx1Q4vvtmZerM0zteegFJV0Xhbqip8V8wAbPYWRhGljqHZdrx8RK7hcM6agMQ588tQ7k3tVs8yqxv0glGGmLp8oUXZUCKxqCEiooSTUuL+n42zNEf0c+I3/ZK745n607ewL/zcEAuceD70wYUWZUSUGEf2d2HW/tmG2MYaDRd/VYGAntqNA4JRXdBsS7mvpjNiUUNERAn30RYvluw1LmW5e1KSdzyrqWy67GzQUPhPPN+ihIgS66qR6bhwWJoh9s1OH+79qTrGiNSgTYjaV7N6CeDrHHuGKIxFDRERJZSmSzwQtZfmd4NcmNQzuffSOF97HEo1l51R1yGEwKMH5uCg3safzadX1eKrHalbBOgD8qB37914WwT8DYUNdSosaoiIKKHeKvFgbWWw8bYAcEeSz9KoP34N+6IvDTH/SRdCHzTUooyIzOFQBV47KhcD0lVD/PffVqTu/hohmi5BYxe0TodFDRERJYxfk/jLUuMszZkFbozslsSzHdWVcL4y2xDSBhci8LvzLEqIyFw9XCqePbwblIjGATvrddz4Q2XKHsypRRU16rIFQIp+LdQ8FjVERJQwrxXVYXNt+N1duwLcNjG5Z2mcrz1uPGRTtcF3+QzAZmthFFHncnAfJ/44NsMQm7vZi38X11uU0b7RRkyAdLoabyuVe6FsLrIwI+poLGqIiCgh6oM6HllWY4hdNCwdQzKTtzhQF38F++KoZWcnXwh9UIFFGRFZ59YJWZjQ3TireuvCKmyqCcYYkcQcTmij9zOEVHZB61RY1BARUUL8c3Uddnn0xttuVeDP4zMtzKhloroCrlejl50NQ+CEcy3KiMhaDlXg2cO7wR1xgE1tUOKqbyoQTME2z8GJxi5otuXcV9OZsKghIqIOt9erYfYK4yzNlSPT0SdNjTHCes5X/w5RU9V4W6q2ULez5J1ZIkq0YTl2PLC/ccnoolI//r6y1qKM2k8bd4DhtrpxHURlmUXZUEdjUUNERB3u4WU1qAmE38nNcQj8cVzyztLYFn0J249fG2L+Uy6CPjDfooyIkselw9Nx3ADjQbmPLK9OuWVoMqc7tPyRhpjKLmidBosaIiLqUMVVAby0ts4Q+/P4TOQ4k/MlR1RXwPna3w0xbcgwBE44x6KMiJKLEAJPHtoNPVzhn2GfBsxYVNXCqOQU3dpZXfWzRZlQR0vOVxgiIkpZ9/9cjWDEcvvBGSquGJkRe4CVpITzldnGZWc2O3xX3AaoXHZG9KtebhX372dchvbxVi8+2Zpah3JqoycbbtvWLAF0PcajKZWwqCEiog6zaLcPczcb/8i5e3IWnBEbjZOJbfGXsP30jSHmP+Vi6APyLMqIKHmdPTQNB/RyGGK3LqqEN5g6TQP0vOGQ7vTG26KmCsq2Egszoo7CooaIiDqElBJ3/Wg8aHNiDztOzXNblFHLRFU5nK9GLTvLG4HAb8+yKCOi5KYIgUcOzDYcyrmpRsOTv9TEHpRsVBu04eONodVLLEqGOhKLGiIi6hBzN3uxeI/fEJs5JRuKSMJZml+XndWGizBps8N7xQwuOyNqwbjuDlw6PN0Q+9uKWmypTZ2mAdroSYbbLGo6hzYXNUKIk4QQ/E1PRERNBHSJ+34ybhqeNtCFQ/s4Y4ywlm3hF7D9/K0h5j/tEsj+Q6xJiCiF3DkpC90jGn94NInbU6hpgDYqqqhZtxwIpk5RRs2LZ6ZmJoCdQoinhBAHtProKEKIaUKIdUKIYiHEjGbudwoh3gzdv0gIMSTivttC8XVCiOPiuOaTQojUa6RORJRiXlpbh5IarfG2IoD7ojYVJwtRWQbna48bYlr+SASmnWlRRkSpJcep4J6on+//bvHi8+2p0TRA758HPbtb423h9UApWWNhRtQR2lzUSCnHAzgGgAfAO6Fi4s7I4iMWIYQK4GkAxwMYBeAcIcSoqIddBqBCSjkUwGwAD4fGjgJwNoDRAKYBeEYIobZ2TSHEfgBy2vr1ERFR+1T5dTy8zLim/sLCNAzPsVuUUQukhPOVv0HURSw7s3PZGVG8zi9Mw+Qexp/xGYuqENRToGmAENBGcglaZxPXnhop5XIp5c0ABgK4FsAZADYIIb4RQpwnhIh1vf0BFEspS6SUfgBvADg56jEnA3gl9PHbAI4WQohQ/A0ppU9KuRFAceh6Ma8ZKngeAXBLPF8fERHF74mVNSjzhVuiptsEbpuYnLM0tgXzYVvyvSHmP/VSyH6DLcqIKDUpQuDRg3IQuWOuqCqIfxfXW5ZTPKKXoNlY1KS8uBsFCCEKANwN4B8AXKGPnwNwHRqKkeb0B7A14va2UKzZx0gpgwCqAHRvYWxL17wOwFwp5c44vjQiIorT9joNT68yrvK9fkwGeqepFmUUm6gsg/P/njDEtIKRCBzPZWdE7TGxhwPnFqYZYg8vrUmJFs/RRY1SvArweSzKhjpCm+fahRDXArgAwFAAcwBcIKVcGHH/OwBKYw1vJhb9f3ysx8SKN1eQSSFEPzTMIB0RIxeDoqKitjysS+D3Irnw+UgufD6ad996B7xa+KWku13iePduFBXtTvjnjus5kRJ5bz0NURdeJqerNqybeg58G3hGRUfgz0hyMev5ODtbYI5wISAb/lzbXq/hoe824bz+yb/xflROTzgr9wAAhBbEri8/Rk3BmIR9Pv6M7JvCwsIW749nAfHxAB4D8EFouZeBlLJeCHFajLHb0LBk7VcDAOyI8ZhtoS5r2QDKWxnbXHwiGgqv4obVa0gTQhSH9uo00do3qKsoKiri9yKJ8PlILnw+mreyPICPSo3vZd01pRvGDx+Q8M8d73Ni+/5TuNYvN8QC0y/HoIMO7+jUuiT+jCQXM5+PQgBXeCrxzKq6xtirO1y46eDeyHIk98khyvgDgK//23h7UNVu+AtPTcjn4s9I4sXzf9tXUsq3ogsaIcRNv34spfw0xtgfARQKIfKEEA40bPyfG/WYuQAuCn08HcAXUkoZip8d6o6Wh4afn8Wxriml/EhK2UdKOURKOQRAfayChoiI2u/en6oMU+7Ds204P2opSjIQlWVwvv6kIaYNHY3AtDMsyoioc7lpXCYy7eGFNeU+HU+tSv7ms03Oq1nFfTWpLJ6i5u4Y8TtbGxjaI3MdgE8ArAEwR0q5SghxvxDipNDDXgDQXQhRDOAmADNCY1ehYbnbagAfA7hWSqnFumYcXw8REbXTl9u9+Hy7zxC7b0oWbEqSHbQpJZwvPWZYdibtDngvvxVQkm/fD1Eq6uFSce3oDEPs6V9qscejxRiRHIJRHdCULUVAxIG8lFpaXX4mhDjq18cKIY6EcY9LPoCapqOaklLOAzAvKnZ3xMdeNOyFaW7sLACz2nLNZh6T0dL9REQUH02XuOsn4wv/oX0cOG6Ay6KMYrN9/ylsy34wxPzTL4fsO8iijIg6p2vHZOC5NXWNnRDrghKPLq/Bwwcm8ekaWTnQBhZA3boBACCkhLpmKbQpv7E4MWqPtszUvBD65wTwYsTt5wFcCuD6hGVHRERJZ06JB7+UBwyxmVOyEdrHmDRExd6my84KxyBw7OkWZUTUeWXaFfx5fKYh9tK6OmyuSe6GAWzt3Hm0WtRIKfOklHkAXv/149C/fCnlwVLK6L0xRETUSfk1ib8sNc7STM93Y2IPh0UZxSAlnC89ClEfXtfPZWdEiXXpiHQMSA//fPl14KFlbVrQY5noooaHcKauNu+pkVJemMhEiIgo+b1WVIctteF18nYFuHNS8h20afv+E9iWLzTE/NOvgOwzMMYIItpXTlXgtonG2Zo3N9RjfWUgxgjracPHQ6rhQkzZtRWiPNYJJZTMWixqhBBrIj7eKoTY0ty/xKdJRERWqw/qeCTqXdeLhqVjSGY8pwMkXsOys6cMsYZlZ7FOHSCijnJ2QRqGZ4d/J+gyyWdr3GnQ80YaQpytSU2tzdRcEfHx+Wg4fLO5f0RE1Mm9sKYOuzx64223KpqsobeclHC+Otu47MzhhPfyGVx2RmQCVRG4PWr29t2NTffhJRO2du4cWixqpJTfRXz8dax/iU+TiIisVO3XMXul8dyJK0amo09achUKtsVfwbbke0PMP/1yyD6JPxCUiBqcONiFMbl2Qyx6L14yCUbvq1m71KJMaF+0eU+NEOImIcSE0McHhpaelQghDkpcekRElAz+sboW5b7wLE2mXeDGsUnWMb+2Co7XHjeEtIJRCEzlsjMiMylC4I6ovTUfbfFi2V5/jBHW0gtGQdrDRZhSvgeibLeFGVF7xHP45h8BbAx9/BcAf0PD2TF/7+ikiIgoeZR7NTz1i3GW5prRGch1JdcsjfP1p6DUVDbeljY7vJfdwmVnRBaYNtCFyT2MszWzliTpbI3d0XRfzfqVFiVD7RVPUZMtpawSQmQCGA/gSSnlCwCGJyY1IiJKBo+vrEVNQDbe7uYUTU4Pt5q6fCHsP3xmiPlPugCy/xBrEiLq4oRourfms+0+LC71WZRRy7TCMYbbStEvFmVC7RVPUbNVCHEwgLMBfCOl1IQQWQC0VsYREVGK2lWv4dk1dYbYH8dmIssRz8tHgnnq4Hz5MUNIG1iAwAnnWpQQEQHAUf2cOKi38QyrWUuSsxOaNmys4TZnalJPPK9KNwN4G8AdAGaGYr8DsLijkyIiouTw2IoaeLTwLE1vt4LLR6ZbmFFTzjf/BaV8T+NtqSjwXXYLYEuuVtNEXY0QArdPNM7WfL3Th293Jt9sjTZ0tOG2sq0EqK+N8WhKRvEcvjlPStlPSjlESvlzKPwWgJMSkxoREVlpc00QL68zztL8eXwm0mzJM0ujrF0G+5dzDbHA8WdBz+PKaKJkcFhfJw7v6zTEHlxaDSlljBEWyciCFrFcVUgJtXiVdflQ3OJ6ZRJCZAsh9hdCHCWEOArAYaF/RETUyfx1eQ0C4YZnGJSh4qJhSTRL4/PC9eIjhpDeewD8p1xsTT5E1KzoTmgLdvvx5Y7km63RC7kELZXF09L5YgA7AHwI4IWIf88nJDMiIrJMUVUA/ymuN8RunZAJhyosyqgpx3svQdm93RDzXnYL4HDGGEFEVjigtxNT+xt/Lh9YknyzNdH7atgsILXEM1MzC8B0KWVvKWVexL/8RCVHRETW+MvSGugRf28UZttwVkGadQlFSduxEfaP3zLE/EefAn34OIsyIqKWRHdCW7I3gI+3ei3KpnnRHdDUkjVAMGhRNhSveIoaG4BPE5UIERElh5XlAby70WOI3T4xEzYlSWZpggEM+u8rEDK8Nk7v3hv+M660MCkiasnEHg6cMMhliD24tAZ6Es3WyJ59oed0b7wt/D4om4sszIjiEU9R8zCAO4UQybNDlIiIOtwDUQfkjc214+Qhbouyacr+33/DXWpcdua7+CbAnTwzSUTU1G1RndBWlgfw4eYkmq0RAnr0bE0R99WkingKlD8CuBNAjRBiS+S/BOVGREQm+7HUj0+iloTcOSkLikiOWRplWwkcc18zxAKHHAdt3AEWZUREbTUm145To94g+cvSamh68szWNDmvhvtqUkY8TfzPT1gWRESUFGZGzdLs39OBYwckycZ7XYPzhUcgtPAadz2rG3znXmthUkQUjxkTM/HBZk/jnr21lUG8u9GDM5Jkz170vhpl/UpASiBJ3tih2Npc1Egpv05kIkREZK2vd3jxTdSheHdOzoJIkhdz+6fvNGzcjeC78A9ARlaMEUSUbIbn2HFGvhtvbgjv23toWTVOzXMnxb49fdBQSKcLwtcwY61UV0CUbofsPcDizKg18bR0dgohZgkhSoQQVaHYsUKI6xKXHhERmUFK2WSW5oh+TQ/Ns4rYvR2Od14wxIL7HQ5tyhHWJERE7XbrhCxEdoffUK3hjQ31sQeYSbVBKxhlDPG8mpQQz56a2QDGADgPwK+LH1cB+H1HJ0VEROb6eKsXP+0JGGJ3TUqSGRAp4XzpUQh/eBYp6EqD74I/WJgUEbVXfpYN5w41Ljf767Ia+LXk2FvT5BBO7qtJCfEUNacCOFdKuQCADgBSyu0A+iciMSIiMocuZZOOZ8cPdGFyT4dFGRnZvv4ItjVLDbHtU8+EjGi9SkSp5eYJmbBH/BW6pVbDi+vqrEsoQpPzajhTkxLiKWr8iNqDI4ToCaCsQzMiIiJTvb/Rg1UV4c33AsAdSTJLI8r3wPnGPwyx4Oj9UD7uYIsyIqKOMCjDhouHpRtif11Wgyq/HmOEebShoyAjTjBRdm4BaiotzIjaIp6i5i0Arwgh8gBACNEXwFMA3khEYkRElHhBXeLBpTWG2On5bozJtVuUUQQp4Xz17xCe8Lu30umC75I/sRMRUSdw84RMpNvCP8vlPh2Pr6xpYYRJ3OnQB+YbQmrRKouSobaKp6i5HcBGACsB5AAoArATwH0JyIuIiEzwn+J6FFeHZ2lUAdw2ITlmaWyLv4Jt6feGmH/6FZA9+1qUERF1pF5uFTeMzTDEnllVi+11mkUZhfG8mtQTT1EzFMBaAA8CuAfAQVLKG6WU/oRkRkRECeXTJB5eZnxX9NyhaSjIjucIswSprYLjtccNIW3oaASOOcWihIgoEa4dnYHe7vCfo14NeHBpdQsjzKFzX03KabWoEQ1eRMMMze0ATgRwBYClQoiXRLIcYEBERHF5ZV0dtkW8I+pQgFsmZFqYUZjz9aehRKxhlzY7vJfeDCiqhVkRUUfLsCu4faJxdvjfRfX4pTwQY4Q5omdqlE3rAL8vxqMpGbRlpuZKAEcAOFBKOVhKeZCUchCAgwAcBuCqBOZHREQJUBfQ8egK4yzNJcPTMTDD+lkadfki2H/41BDzn3QBZP8h1iRERAl1XmEahkfMEEsA9/5UZV1CAGRuL+jdezfeFsFAQ2FDSastRc0FAG6QUv4YGQzdvjF0PxERpZDn1tSh1BPuMpRmE/jT+CSYpfHUw/nyY4aQNrAAgRPOsSghIko0myJw737G2Zr52334aofXoowaNGntvGGNRZlQW7SlqBkF4OsY930dup+IiFJEpU/H7KgOQ1eNTEcvt/VLuxxvPQulvLTxthQKfJfdDNiSoBsbESXMtIEuHNzbeDbW3T9WQ5fWHcipF4w03FZY1CS1thQ1qpSy2f56oXg8zQaIiMhis1fUoMof/kMhyyFww1jrZ2mUdSvg+Px9Qyxw/JnQ80ZYlBERmUUIgZlTsg2xFeUBvF3isSgjQMs3FjVqCYuaZNaWxdN2IcSRaDiPrb3XICKiJLC9TsO/1tQaYjeOzUQ3p8XvT/l9cL34iCGk9+4P/ykXW5MPEZluck8HTh3ixnubwoXMzCXVOGmwGy6b+X2p9EFDIVUbhNbQ9l4p2w1RWQaZ0930XKh1bSlISgG82Mr9RESUAh5aWg1vxBEQfdwKrh6VHnuASRwfvApl11ZDzHfJnwGny6KMiMgKd0/Own+3eBAIbfnbWqvhuTW1uN6K2WSHE/qgoVA3rm0MKSVroU06xPxcqFWtvjUnpRwipcxr6Z8ZiRIR0b5ZWxnA68X1hthtE7OQZrN2lkbZXAT7vP8YYoEjT4Q2cqJFGRGRVfKybLhshPGNlkdX1KDCp8cYkVhaAZegpQruhyEi6iLu/7kaesSe28JsG84rTLMuIQAIBuF8/mEIPfwHi96tB3xn8rQAoq7q5vGZyLKHl5tV+SUeW97s9u6E0/OjmwWstiQPap1pRY0QYpoQYp0QolgIMaOZ+51CiDdD9y8SQgyJuO+2UHydEOK41q4phHhBCLFcCLFCCPG2ECIj0V8fEVEyW7jbh3lbjO1R756cBZti7fnJ9v+9AXVLsSHmu+gmII2/tom6qu4uFX8cZ1xu9uyaWmyuCZqei1ZgbPKrblwH6NbMGlHLTClqhBAqgKcBHI+GFtDnCCGiW0FfBqBCSjkUwGwAD4fGjgJwNoDRAKYBeEYIobZyzT9KKcdLKccB2ALguoR+gURESUxKiXt/qjbEpvS043eDrN2vomzZAMd7LxtigQOPhjbxYGsSIqKkcfWoDPRPC7eZ9+vArCXVLYxIDNm7P2R6+Awd4amD2LnF9DyodWbN1OwPoFhKWSKl9AN4A8DJUY85GcAroY/fBnC0EEKE4m9IKX1Syo0AikPXi3lNKWU1AITGu9FwOC0RUZf08VYvFpb6DbF79stGw69IiwQDcD73YGNXIQCQmdnwnXe9dTkRUdJw2wRun2ScrZlT4sGyvf4YIxJECGj5xrbyPIQzOZlV1PQHENnWZlso1uxjpJRBAFUAurcwtsVrCiFeArALwAgAT3bEF0FElGqCetNZmuMGOHFoH6dFGTVwfPAq1C0bDDHvxX8CsnIsyoiIks3ZBWkY3c3YqPeuH6sgTT6QM/oQTrWE+2qSkVlnzDT3dmD0/5GxHhMr3lxB1nhNKeUloSVqTwI4C8BLzSVWVFTUXLhL4vciufD5SC6p+nzM2WHDuqrwKd0CEhf3rERRUYVlOaVt34hhH75uiJWP3h+bs/sCcXyfU/U56az4fCSXzvJ8XNVPwQ0V4aWy3+7y45XFJTgk17x9LVmubBRE3A6sXtau729neU6sUlhY2OL9ZhU12wAMjLg9AMCOGI/ZJoSwAcgGUN7K2BavKaXUhBBvArgZMYqa1r5BXUVRURG/F0mEz0dySdXno8Kn4/kfdyHyPaRzhqbj+AkDrEvK70PaCzMhZES3s5zucFxzJwozsloYaJSqz0lnxecjuXSm3CG+MwAAIABJREFU52OolHi3ogxf7fA1xp7dkYELpvSCalajk769gDeeaLzp3rMdhYMGxnWOVmd6TpKVWcvPfgRQKITIE0I40LDxf27UY+YCuCj08XQAX8iG+cW5AM4OdUfLA1AIYHGsa4oGQ4HGPTUnAlgLIqIu5qGl1ajwhQuaDJvA3ZPbXjgkguOdF6BEbbL1XXozEEdBQ0RdhxAC9+2XZVi2s7oy2OTMrYTKyIbeO7xrQug6lE3rzfv81CamFDWhPTLXAfgEwBoAc6SUq4QQ9wshTgo97AUA3YUQxQBuAjAjNHYVgDkAVgP4GMC1Ukot1jXRsFztFSHESgArAfQFcL8ZXycRUbJYVxnA82vrDLGbxmeiT0Q3IbMpa5fD/slbhljgNydAG3+gRRkRUSoY392BMwvchtiDS6pRFzBvCZoWdV6NyvNqko5Zy88gpZwHYF5U7O6Ij70AzogxdhaAWW28pg7gkA5ImYgoZd25uApaxM7FQRkqrhll4dkvnjq4nn8IImKDr96jN3znXGNdTkSUMu6clIX3N3ng0xpu7/LoeHpVLW6ZYM4sr14wClgwv/G2WrIGAVM+M7WVaYdvEhGROT7b5sVn232G2Mwp2XDZrGvh7HztCSh7dhpivstnAO50izIiolQyMMOG30e9MfPEylqUejRTPn/0TI3Cts5Jh0UNEVEnEtAl7lhcZYgd3NuBkwZbd9CmbdGXsH//iSHmn3o6tJETLcqIiFLRH8dlItcZ/tO1Nijx0NIaUz63PqgA0mZvvK2Ul0JUlpnyualtWNQQEXUiL6ytw/qq8IGWAsBfDrDuoE1RVgrny48ZYlr/IfCfeaUl+RBR6sp2KLhlgvFAzlfW12F9pQkLwewO6IOGGkJKCWdrkgmLGiKiTqLUo+EvS40HbZ5fmIbx3R0xRiSYrsH57IMQ9bWNIWmzw3f1XYDD2sM/iSg1XTo8HfmZ4YYnmgTuiTpgOFG06EM4uQQtqbCoISLqJO5cXIUqf3gjfqZd4C4LWzjb/zcHtrXLDDH/GVdCH1QQYwQRUcscqsA9+2UbYv/b6sX3u3wxRnQcPXpfDWdqkgqLGiKiTuDrHV7MKfEYYrdMyEQvtzUtnJVN6+F45wVDLDh6MgLHnm5JPkTUeZw02IX9expnoO/6sQp6RHfFRGgyU1OyFtDNaVRArWNRQ0SU4rxBiZsWVBpiY3LtTToFmcbnheufD0Bo4b09Mj0LvituAxS+7BDRvhFCYOYU4yz0kr0BvLfRE2NEx5C9+kNGHBQsvPVQdmxpYQSZia8uREQp7u8ra7ChOvxuoQAw+6Ac2BRrmgM4X38Syk7jC7330pshu/WwJB8i6nwO6O3EiVFdHe/7uRo+LYGzNUI0be3MJWhJg0UNEVEK21AVxN9WGFuaXjw8DVN6WdMcwLbwc9i//sgQCxz+W2j7HWZJPkTUed07ORuRx29tqdXw3Jra2AM6QHRRw2YByYNFDRFRipJS4k8LK+HXw7GeLgX3TM6OPSiBROkOOF8ytm/W+w6E77zrLMmHiDq3gmwbLhlhPMD30eU1qPTpMUbsO72AMzXJikUNEVGKervEg692GDv+zNo/GzlOC361BwNwPXMfhLe+MSTtdnh/fzfgSjM/HyLqEm6dkIkse3i6ptIv8ejyxB3IqeWPMNxWtpUA/sR3XqPWsaghIkpB5V4Nty+uMsQO7+vEGfluS/JxvPUc1I3rDDH/2ddAH1xoST5E1DX0cKm4cZzxQM5n19RiU00wxoh9lJENvVe/xptC16FsLkrM56K4sKghIkpBf15YhT3e8BILhwI8dlA2hDC/OYC6fCEcH88xxIKTDkXg6FNMz4WIup7fj8pA/7Rw+3q/Dsz8OXEHcjbZV8MlaEmBRQ0RUYp5f6MH70a1Lr1pXCYKs+2m5yIq9sL13F8MMb17b3gvuwWwoMAioq7HbRO4Y5JxtuadjR4s2eNPyOfTo5eglaxNyOeh+LCoISJKIaUercmZNONy7fjT+MwYIxJIC8L1z5kQNeFlcFJR4L36TiAjq4WBREQd66yCNIzJNb6xc/dPVZAJOJBTyzMWNZypSQ4saoiIUoSUEn/8oRLlEZ197Arwj8O6wW7BmTSO916Guna5IeY/7VLow8aangsRdW2qIjBzP+ObKd/t8uPz7R2/iV8fXAgZcZCwUroDqE3ccjdqGxY1REQpYk6JBx9t8Rpit03Mwuhc85edqSsXw/7f1w2x4KhJCJxwjum5EBEBwJH9XTiyn9MQu+enKmh6B8/WOF3QB+QbQupGLkGzGosaIqIUsKNOwy0LjcvOJvew44YxGabnIspL4frXLIiIZR16di58V98JKGoLI4mIEuveqNmaVRVBvFXiifHo9tPzo8+rYVFjNRY1RERJTkqJG3+oQJU/XEQ41YZlZzazl51pQbj+EbWPRijwXnM3ZHauubkQEUUZ393RpLX9A0uq4Q127GxN9Hk13FdjPRY1RERJ7p+r6/DpNuO68DsnZWFYjvnLzhzvvAB1/UpDzH/aJdBHTDA9FyKi5twxKQv2iL9wt9VpeGFdXYd+Dj2vmQ5oCWhKQG3HooaIKIkt3evH3T8ZD9k8sJcD14wyf9mZumwBHB/9xxALjp2CwO/OMz0XIqJYhmTacOnwdEPs0eXVqPLrMUbET+8/GNLharytVFdAlJd22PUpfixqiIiSVJVfxyVflSMQ8TqcZRf45+HdoJq87EyU7Ybr2ajzaLr1gPfKOwCFLyVElFxunpCJTHv492SFT+KJlTUd9wlUG/QhwwwhhUvQLMVXIiKiJCSlxI3fV2JTjWaIP3loNwzJtJmbTDAI1zP3Q9SFW5ZKpWEfDbJyzM2FiKgNerhUXB/VSOWZVXXYWa/FGBG/pvtq2CzASixqiIiS0Mvr6vHeJmPHnstHpOPkIe4YIxLH8dazUItXGWL+6ZdDHzbO9FyIiNrq2tEZ6OUO/6nr0ST+uqzjzpOJ7oDGZgHWYlFDRJRkfikPYMZiY/vmMbl2PDAl2/Rc1CXfw/HxHEMsOP5ABI4/2/RciIjikW5XMGOCscXza+vrsakm2CHXj56pUTauA/SOmwmi+LCoISJKItWhfTS+iNfFdJvAy0d0g8tm8j6aPTvhei5qH01uL3ivvI37aIgoJVwwLA15meHzs4ISeHR5x+ytkT36QGaG32wSPi+UHVs65NoUP74qERElCV1KXPVNBYqqjO8izj44B0OzTW7fHAw07KOpr20MSVWF99p7gAzzZ4yIiNrDrgjcEjVb85/iepRUd8BsjRDQeAhn0mBRQ0SUJB5eVoP/bfUaYucVpuHMgjTTc3G8+c8m68P9Z1wJfeho03MhItoXZ+S7MTQr3GBFk8DDHbS3RsvjIZzJgkUNEVES+HCzBw8vMy6JGJdrxyMHWrCP5qdv4Pj0HUMsOPEQBKadaXouRET7yqYIzJiYaYi9VeLB+srAPl9bj95Xw5kay7CoISKy2JqKAH7/TYUh1sOl4PWjc5FmM/fXtNi9Da7nHzbE9B694b1iBiDM3dNDRNRRTh3ixoic8GyNLtHkjaT2iJ6pUbZtAPy+fb4uxY9FDRGRhSp9Os77vAy1QdkYswnglSNzMTDD5PNo/D64nroHwlPXGJKqDd5r7gXSM2OPIyJKcqoimnRCe3ejB6sr9nG2JisHes++jTeFpkHZUrxv16R2YVFDRGQRTZe47OtylEQdsPnQAdk4pI/T9Hycrz0OdcsGQ8x/zjXQC0bGGEFElDpOGuLC6G7hN4skgIeW7vveGh7CmRxY1BARWeT2xVX4fLtxmcKFw9Jw2Yh003Oxffcx7N/MM8QC+x+JwDGnmp4LEVEiKELgtonG2Zq5m71YUebfp+tGH8KpsFmAJVjUEBFZ4Lk1tfjXmjpDbP+eDjxyYA6EyXtXlK0lcL4y2xDT+wyE79KbuY+GiDqVEwa5ML67sUX+X/dxb03TDmicqbECixoiIpPN3+bFrYuqDLH+aSpePSoXTtXkIsJT37CPJmJjq3Q44b3uPsBtfitpIqJEEkLg9qjZmo+2eFFc1f69NfqQQkgR/pNa2b0NqK1qYQQlgmlFjRBimhBinRCiWAgxo5n7nUKIN0P3LxJCDIm477ZQfJ0Q4rjWrimEeD0U/0UI8aIQwuRT64iImre6IoBLviqHHu4LgAybwBtTu6NPmhp7YCJICeeLj0DZtdUQ9l30R+gD883NhYjIJMcOcGJcbvhPQwngmVV1sQe0xumGPjDPEFI3cAma2UwpaoQQKoCnARwPYBSAc4QQo6IedhmACinlUACzATwcGjsKwNkARgOYBuAZIYTayjVfBzACwFgAbgCXJ/DLIyJqk931Gs78rAw1gXBFowjg+SO6YWyu+e+92Oe/B/viLw2xwOG/RfDQaabnQkRkFiEErh+TYYj9u7gOe71ajBGt0wuMBxOrG1a3+1rUPmbN1OwPoFhKWSKl9AN4A8DJUY85GcAroY/fBnC0aFhYfjKAN6SUPinlRgDFoevFvKaUcp4MAbAYwIAEf31ERC3yBCXO+6IM2+qML5oPTMnGtIFu0/NRNqyB4z/PGGLaoAL4LviD6bkQEZntlDw3BqSHZ8e9GvD8mvbP1mhDjUWNUryq3dei9jGrqOkPIHJ9w7ZQrNnHSCmDAKoAdG9hbKvXDC07uwDAx/v8FRARtZOmS1zxdTl+2mNcs33ZiHT8fpT5nc5QWwXX0/dCaMHGkHSnw3vtfYDD/FbSRERmsysCV0f9/n1uTR08EWeGxUMbalyApG5YA+jtn/mh+Jl1sltzO1+j/6+J9ZhY8eYKsuhrPgPgGynlt7ESKyoqinVXl8PvRXLh85Fc2vt8SAk8VmLHf3cal5cdmKPh8u57UFy8pyPSiyMhHflvPgWlbLchvPGEC1FV4wFqUuf/O/6MJBc+H8mFz0frDlWBDNWNWq3hT80yn44nftiE0/oGWxnZDCkx1p0Bm6cWACC89di64Ft4e4Xfb+dzsm8KCwtbvN+somYbgIERtwcA2BHjMduEEDYA2QDKWxkb85pCiHsA9ARwVUuJtfYN6iqKior4vUgifD6Sy748H0/9UoM3dxoPdxuZY8ObJ/RFtsP8BpT2D1+Hs3ilIeY/djp6nXgWepmeTfvxZyS58PlILnw+2u6y2io8/ktt4+05e9y4+bDeUNrTzn7YGGD5wsabeYFaBEPPA5+TxDPrFfVHAIVCiDwhhAMNG//nRj1mLoCLQh9PB/BFaE/MXABnh7qj5QEoRMM+mZjXFEJcDuA4AOdIKfUEf21ERM16b2M97vzRWND0TVPw1tTulhQ06pqlcLzzgiGmFYyC/6wW3/shIuq0rhqVAVtE/bKhWsO8Ld52XUsriFqCxn01pjLlVTW0R+Y6AJ8AWANgjpRylRDifiHESaGHvQCguxCiGMBNAGaExq4CMAfAajTsjblWSqnFumboWv8E0BvAAiHEMiHE3WZ8nUREv/phlw9XfVNhiGXaBeZM7YEBGWZNkoeJir1w/uN+iIj3eWRGFrzX3gvY2PWeiLqmfukqpucbm7U8FTFzEw89qlkAixpzmfbKKqWcB2BeVOzuiI+9AM6IMXYWgFltuWYobv5fDEREIesqAzj38zL4I+aJbQJ49chcS1o3I+CH68m7oVSFiywpBLxX3wnZPZUWnRERdbzrxmTijQ2extsLS/1YXOrD/r3ia5yi5Y+EFAJCNmzxVnZuAepqgPTMDs2Xmmf++gciok5sd72G6Z+VodJv7FvyxCE5OLK/y/yEpITztcebnJkQOOkCaGP3Nz8fIqIkMybXjqP6GQuYds3WuNOg948+hJPn1ZiFRQ0RUQepDeg4c34ZttYa23jePjET5xZa0LoZgO3LubB//ZEhFhx3APynXBRjBBFR1xN9GOdHW7zYXR9/S+amS9BY1JiFRQ0RUQcI6hKXfFmO5WXGs2guHJaGm8dbs/RAWb8Szv970hDTe/eH9+o7AUWNMYqIqOs5op8TI3LCuxc0CfynuD7u60SfV6NwpsY0LGqIiPaRlBI3LajEZ9t9hvgx/Z147KAciPa0Bt1HonwPXE/dbTxg0+mC94aZXN9NRBRFCIELhhln1F9dXwcp4zuMs0kHtJLVgM5GvGZgUUNEtI8eXV6DV9cb39Eb392Ol4/MhV0xv6BBwA/XU8bGAADgveI26APyzc+HiCgFnF3gRmS3/ZIaDd/v9sd1DdlnIGTEG0eivg7Kjs0dlSK1gEUNEdE+eHV9HWYtrTHEBmaomHNMd2TYLfgVKyWcLz4KdcMaQ9h/4vnQpvzG/HyIiFJEd5eK3w02tnd+dV1dfBdRlCazNVyCZg4WNURE7fRuST3+8H2lIZbjEHh7anf0TrNmz4rj/Zdh/+FTQyw47gD4T7vEknyIiFLJhcPSDLc/2OxBpS++5WM8hNMaLGqIiNrhf1s8uPKbCkSutnYowOtHd8fwHGsOs7R99zEc779iiOm9B7AxABFRGx3e14lBGeHflz4NmLMhvoYB0R3QFHZAMwWLGiKiOH29w4uLvypHMKKiUQXw4hG5OKRPfIe1dRR11c9wvviIISYzsuD500NsDEBE1EaKELig0Dhb80qcDQO0/BGQEQ1i1B2boHrj76RG8WFRQ0QUh8WlPpz7eTl8EccXCAD/OKxbk7XYZlG2bQx1OgsnJe12eG58ELL3AEtyIiJKVecWpiOyx8uqiiCWRbXrb1FaBvR+g42h7Rs7KDuKhUUNEVEbLdvrxxmflaEuaHzHbvbBOTizIC3GqMQSlWVw/W0GRL1xM6v3yjugF46xJCciolTWP13F1P7GWfdX18fXMCB6CVr69g37nBe1jEUNEVEb/Fjqx0mf7EWV31jQzJyShYuHp8cYlWB1NXA9diuUst2GsO+sq6Htf4Q1ORERdQLRZ9a8XeJBbaDtDQOimwWkc6Ym4VjUEBG14ucqBad+shfVUQXNrRMycf0Yi/ar1NfC/cjNULcUG8KBI09C4PizrMmJiKiTOG6gC73c4T+TawIS72/ytHm8FjVTk7a9hIdwJhiLGiKiFnyx3Ys/rHKiNmrJ2R/GZGDGBIsKGk893I/dCnXjWkM4OO4A+C64ARAWHPhJRNSJ2BWBc4calxW/tr7tm/1l30GQaeHZHpu3HmLX1g7Lj5piUUNEFMP/tnhw9vwy+HRjkXDrhEzcu18WhBXFg88D9+zbmpx7oA0fD+919wGqzfyciIg6oQsKjUvQFpX6sakm2LbBigItP+q8mqJfOio1agaLGiKiZrxdUo8LviiHP2q1wL2Ts3DbRIsKGr8Prr/fAXXdckNYGzoGnj/+BXC6zM+JiKiTKsi24cBeDkPs3Y1xLEGLatairl3WIXlR81jUEBFFkFLi7ytqcPnXFYhacYaHD8jGjeMsWnLm98H1xF2wrV5iCGt5IxrOonFb032NiKgzOz3f2Kr/nZK2L0HTRkww3FbXLAXiOO+G4sOihogoJKhL/HlhFe79udoQF5B44pAcXDUqw5rEaqvhfvhPsK1cbAhrgwvhufkRIM2ivIiIOrmTh7ibnFmztrJtZ9boBSMh7eGZHqViL0Tp9o5OkUJY1BARAagL6Dj/i3K8sNZ4FoFdAWYO9+PCYda0bRZ7dyHtgeugFhvXYmsD8hsKmnSLZo6IiLqAXm4Vv+lrPLPmnZI2LkGzO5ouQVvDJWiJwqKGiLq8PR4NJ368Fx9v9RriWQ6Bd47tgeN6apbkpWzZAPfM66Ds3GKIawPy4b31MSAzx5K8iIi6ktPyjEvQ3t1YD9nGZWTNLkGjhGBRQ0Rd2sryAI78cA+W7DUuJxiQruLj3/bE4VHv0JlFXbMU7gdvgFK51xAPjpgAz+2PQ2Z1syQvIqKu5sTBbtgj/mLeUK1heVnblqBpIycabqtrl3FfTYKwqCGiLuvDzR5M+2gPttUZZ2LG5Nrx6Qk9Maqb3ZK8bN9/Ctejt0B4jEvhAvsfCe+f/8olZ0REJspxKjimv7G75Dtt7IKm54+AdITHKpVlPK8mQVjUEFGXI6XEI8uqccEX5aiLanF2ZD8n5h3fA/3SVfMTCwbh+L8n4Hr2QYig8V1A/9TT4fv9XYDdEWMwERElSnQXtPc2eqC3ZcbFZm9mXw2XoCUCixoi6lI8QYnLv67ArKU1Te67YkQ65kztjiyH+b8aRWUZ3A/fBMdn7za5z3fmVfCfdx2g8Fc2EZEVjh/oQpot3AZtW52GxaX+No1tdgkadTi+QhJRl7G5Johp8/Y0WTZgE8Dsg3LwyEE5sCvmH6qpFK+C+56roK5fYYhLuwPeq+5A4IRzACsO+yQiIgBAul3BtIFRS9Da2AVNGxnVLID7ahKCRQ0RdQlfbvfiyA/3NNnc2c0p8N5xPXDJCAtaNksJ2xcfwP3gH5o0BNB79IbnzqcQPHiq+XkREVETp0d1QXt/kwdBvfXiRB8yHJo93HRGqaqA2LG5w/Pr6ljUEFGnJqXE7BU1OP2zMpT7dMN9I3Js+PLEXjjMig5ndTVwPn0fXK/MhtCChruCoyej/t5/QR8yzPy8iIioWccMcCHLEZ413+PV8d0uX+sDbTbUDSo0hLgEreOxqCGiTqsmoOPCL8tx38/ViH4z7YRBLnx6Qk8MybSZnpdSvAppd18B+49fNbnP/9tz4P3TwzyDhogoyThVgd8NMs7WvN3GJWg1g4cbbtvYLKDDsaghok5peZkfR87dgw83Gw/UFADunpyF147KNb8hgK7D/tG/G86f2bvLcJd0uuC95h74z7oKUM0vtIiIqHXTo7qgfbjZA7/W+hK02qiiRl27DND1GI+m9uArJxF1KlJKPLumDnf9WAV/1OtFjkPghSNycXTUeQNmEJVlcD73EGy//NjkPm1wIbzX3A3ZZ6DpeRERUdsd3teJHi4Fe70NLzBVfomvd/owdUDLryv1fQdButIgvPUAAFFTBWXHJugD8hOec1fBmRoi6jQqfDrO/6Icty5qWtCMzbXjq5N6WVLQqEu+R9qdlzZb0PiPPR2eu55mQUNElAJsisCJg42vIx9sasMSNEWFNnycIaSu4b6ajsSihog6hR92+XDYB6X4aIu3yX0XDUvDJyf0MH//jM8D58uPwf34HRA1VYa7ZHoWPH+YBf951/NATSKiFHLyEOMStI+2eBBoQxe0JufVcF9Nh+LyMyJKabUBHff9XI3n19Qh+iUlyy7w94NzcFp+mul5KRvXwfWvB6Ds3NrkPm34eHivvgMyt5fpeRER0b45pI8TuU6lsaNmhU/i250+HNXKSgBtRPR5Ncsb9tXwYOUOwaKGiFLWF9u9+MMPldhaqzW5b1IPO148Itf82Rldg/1/b8LxzgsQmjEvqarwn3opAiecDSiquXkREVGHsCsCJwxy4bWi+sbYB5s8rRY1+uChkGkZEPW1AABRVw1lWwn0QUMTmm9XwdKQiFJOpU/Htd9V4LRPy5otaK4fk4GPf2t+u2ZRVgr3wzfBOefZJgWN3nsAPHc9jcCJ57GgISJKcdFL0P672dv6QZyKCm34eEOIS9A6DosaIkoZAV3iX6trMemd3Xg94h2yX+Vlqpg7rQdmTsmGQxXNXCFxbIu+QNqdlzYsJ4gS+M3vUD/zOeh5I0zNiYiIEuPwvk5kRxzEWebT8f0uf6vjmixBW82ipqOYVtQIIaYJIdYJIYqFEDOaud8phHgzdP8iIcSQiPtuC8XXCSGOa+2aQojrQjEphOiR6K+NiBJLSokPN3tw4Hu7ceuiqsZ1zL9SBHDt6Ax8f0ovHN7XaW5ynjo4n/0LXM/c37ik4FcyIwueG2bCd+mfAac7xgWIiCjVOFSB30YdxDl3c+td0LRRUc0CVv8M+Jo2uKH4mVLUCCFUAE8DOB7AKADnCCFGRT3sMgAVUsqhAGYDeDg0dhSAswGMBjANwDNCCLWVa34P4BgAmxP6hRFRwi3a7cNv/7cXF3xRjg3VTZeajcix4dMTemLW/tlIs5k7+awUr0LaXZfD/v0nTe4Ljt4P9Q+8CG3yYabmRERE5jhlSNODOLVWlqDpAwugRzSJEX4f1Gba/VP8zFpwvj+AYillCQAIId4AcDKA1RGPORnAvaGP3wbwlBBChOJvSCl9ADYKIYpD10Osa0opl4ZiCf2iiCgxpJT4dJsPf19ZgwW7m5/OT7cJ3Dg2AzeMzYTT5KVm0IJwzH0N9rmvQUSdCC1tdvjPvBKBqaezow0RUSd2RD8nsuwC1YGGQqbUo2NhqR+H9GlhxYAQCE4+FI7P3m0M2ZZ8xzfAOoBZRU1/AJF9TbcBOCDWY6SUQSFEFYDuofjCqLH9Qx+3dk0iSiEBXeKdEg+eWFmD1ZXBZh+jiIZzZ2ZMyELvNPM33IvSHXD9axbU4lVN7tP6D4Hv6rugDyowPS8iIjKXUxWYNsiFORvCy84+2ORpuagBGgqYyKJm6QL4tCCgsinxvjDru9fc26jR83OxHhMr3txboK2ffBSlqKgo3iGdFr8XyaUrPR+7vALv77Zh7m4Ve/yxZzcO7abh+jw/8tPqUb19L6pNzLFo/XrkrliAAZ/8G6rf1+T+0ilHY8dRp0H6dKALPXdW6ko/I6mAz0dy4fNhjv0dKuYgXMS8t6EGl+XugdLMX6+Nz4nixlh3OmyeOgANrZ13fj4PtXkjzUg5ZRUWFrZ4v1lFzTYAAyNuDwCwI8ZjtgkhbACyAZS3Mra1a7aqtW9QV1FUVMTvRRLpCs+Hpkt8us2Ll9fV4bPtPrS0DPnIfk7cNC4Th5ndBCBk49KfMPKb92Bb8n2T+/TsbvBdNgNp4w8ATxowT1f4GUklfD6SC58P8wzIk7i3aCdqgw0vYnv8CiqzB+GA3sbXq+jnRE4+FPguvB9zcOkm+I89yZykOymzipofARQKIfIAbEfDxv9zox4zF8BFABYAmA7gCymlFELMBfBvIcTfAPQDUAhgMRpmcFq7JhElESkllpcj3UduAAAgAElEQVQF8FaJB+9urMfOej3mYxXRsAnzhjEZmNDDYWKWEaSEbcF8jHxlNmzepi2kgxMOhu+ymyGzulmQHBERWc1tEzhuoAvvbIxYgrbZ06SoiRacdCjsEUWN7efv4D/veoD7wdvNlKImtEfmOgCfAFABvCilXCWEuB/AT1LKuQBeAPBaqBFAORqKFIQeNwcNTQWCAK6VUmpAQ+vm6GuG4jcAuAVAHwArhBDzpJSXm/G1ElFTm2qCeGtDPd4q8WB9VfN7ZX6VbhM4e2garhudgbws69YXi8oyOF/+G2xLm87OSIcTvnOuQfDIk/gCRETUxZ00xG0oauZu8uKBKRJKC68P2pgpkA4nRGg5s1JeCmXTeuh5wxOeb2dl2l8MUsp5AOZFxe6O+NgL4IwYY2cBmNWWa4biTwB4Yh9TJqJ9UObV8N5GD94q8WBRaesHko3NteOS4emYnu9GlsPCrmG6Dtt3n8D5xjMQdTVN7tYKRsF7+a2Q/QZbkBwRESWbqQOcSLMJ1IeWoG2r0/Bjqb/l2RqnC9rY/WH7+dvGkO3nb+FnUdNubLNARB2myq/j061evF1Sj8+3+xBspXVHhk3glDw3Lhmejkk97Ja3YVc2roXztcehbljT5D5pt8N/2mUITDsDUMzvukZERMkpzabg2AEuvL8pPFvzdknblqBFFjXqz98B07mwqL1Y1BDRPtnj0TBvixf/3ezBVzt9CMTeJgMAsAng6AEunJnvxvGDXKYfmNmsmko433oetm8+gpBNK7G6/vnAdfdwdoaIiJo1Pd9tKGre2+TBgwdkw95cG7SQ4ISDIBWl8bwzdccmiF1bIfsMjDmGYmNRQ0RxkVJiRXkA87f5MH+7F4tK/S12LvvVgb0cOKPAjVOGuNHdlSQzHX4f7F/9F473X252qZm0O+A/7VKsL5iIQhY0REQUw9QBLmQ7BKr8DS+Ie706vtrhw9QBrtiDMrKgjZgA2+oljSHbz98hcMI5iU63U2JRQ0St2lmv4YddPszf7sMX273Y7WllOiZkRI4NZ+SnYXq+G4Mzk+jXjc8D+5cfwj7vDShV5c0+JDj5MPjOuQayZ1+eO0NERC1yqgInD3Hj1fXhTplvldS3XNQA0CYdaixqlrCoaa8k+iuDiJKBlBIbazQs2O3DD7v9+GGXDxtrtDaPL8hSceJgN07Lc2NsrvX7ZAw8dbDPfx+OT+ZA1FQ1+xC970D4zrsB2tgpJidHRESp7Iz8NENR89FmL+qDeovLrIOTDoXz/8K9rdTiVRCVZZA53ROaa2fEooaoC9OlxJZaDcvLAli6149lof/+On3eVuNy7fjdYBdOHOzGiBxbchUyug5l/QrYf5gP249fQtTXNfsw6XLDf/JFCBx7OmCzm5wkERGluoN7O9AvTcGO0BlsdUGJ/23x4vT8tJhjZPde0PKGQ924rjGmLvkOwaNOTni+nQ2LGqIuQEqJXR4d6ysDWF0RxJrKANZUBLCmIth4CnI8MmwCv+nnxDH9XTh6gBODMpLsV4mUULZvhO2H+bAtmA+lvDT2Qx0uBI46CYHjz+I7Y0RE1G6qInBaXhqeWlXbGHurxNNiUQM0LHeOLGpsLGraJcn+EiGifeENSmyoDqK4Ooj1lQEUVQdRVBVEcVUQNYH4i5df2QQwqYcDB/dx4Oj+LhzQywGHmkSzMcEAlM1FUItWQS3+BUrRKiiVe1scIl1pCEw9Df7jpgOZOSYlSkREndkZBW5DUTN/mxfl3paXcAcnHwbn28833lZXLwVqq4GMrITl2RmxqCFKEVJKVPoldtRp2FmvYUd9w3931mnYXqehqDqIzTUa2l+6hGXZBSb0cODg3g4c1NuJKb3sHdN6OeAH/D6IYADQgkAwCAQDje0sAUCGlq4JKRseE/A3PCYQAAJ+iJpKiIq9UMr3QFTsgSjfA2X3NohA6wd8AoCemYPA0ac0LDNLz9z3r4mIiChkXK4dw7JtWF8VBAAEJfDBJi8ObaHpp+w7CHqfgVB2bQUACC0I+zfzEPjt2Wak3GmwqCFKAn5NNhQooX8rttsQqKhqLF52heKtvNnTLtkOgdHd7JjYw4GJPeyY0N2O/CwblHj3xUjZUGxs3whl+yYou7ZBVFeE/1VVQHjrW79OAkiHE8FJhyJ48FRoo/cDbPzVR0REHU8IgTPy3Zi1NHxMwFsl9Ti0sMVBCBw2Dc63nmsM2ee/h8Bx0wGVr1dtxe8UUQJJKVHh07GjXm8sWH6daWkoWHTsqtew1xvdItkBoLa5S7Zbpl2gMNuGYdk2jO5mx8hudozqZkffNKV9G/urK6GuXdbwb0sxlO0bY27Ct4K0O6ANH4/gQccgOPkwwN3ymmYiIqKOMD0/zVDU/LDbj10DBVqqawK/OQGO919pXHWglO2GunQBtP0OS3C2nQeLGqJ9JKXEznodqyoC2FAdxOaaIDbVaA3/rdVQ346N+O0lAAzMUFGYbQsVMHYMDRUyvd3tLF5+VVsFde1yqGuWNhQy2zZ2WN4dQe/WA1rhWOiFo6ENHQN9UAG7mBERkenysmzYr6cdP+0JNMY+3auixfIkMwfBg46B/Zt5jSH7Z++wqIkDixqiOEgpUVKtYWGpDyvLA1hVHsCqiiDKfW07jHJfpdsE+qap6JumoG+6in5paui2ivwsGwqybHDbOmgDf10N1HUrQkXMUihbSxr2uewDqSiAKw1StTUsAbPZG6bWFQUSAiJyR5CUkHY7YHMANjuk3QHY7ZDudMjcXpC5PaHn9oTs1hOyey/IrG77+AUTERF1jDPy0/DTnvB5aB/vsWFmK2MCU08zFDW2tcugbNnQ8CYdtYpFDVELdCmxoiyABbv9WLDbh4WlfpR6Or6AUQTQy9VQqPRNU5EWqMHIvt3RN01Bv3QVfUKFS5ZdJO4MGE9dRBGzDMrmYggZ39cqXW7o/YdA758Hvd9gyNyekFndoGfnQmblAGmZgNIBDQeIiIiS2Kl5bty2uAp66L26ojoFK8r8GNfdEXOMPmgotBHjoa5d3hizz38XvktvTnS6nQKLGqIou+o1fLHdiy92+PDldh/K9nEWJt0mGgoTd9PZlX6hIqa3W4FNCRcrRUXlKCxMYGcuKSHKS6GUrIVasrahiNm0ztCFrNVLCAF9UCG0kROgDR8PfWA+ZPfeLFqIiKjL6+VWcURfJ77Y4WuM/WN1Hf5xWOyiBgD8U0/H/7d33/FxVWfCx3/PjKRRlyzZlm3Zxt0yBpcYsGEdCNWmk4ROaAE2b4B3311eskvZEF7KZskSAg4hkIQSSkINzRRjSABTTBIbV2zhCsbYki2r2iqjmef94x5JI1kjaRSNpJGf7+dzP3Pn3HvPLWeuNM+cc89Jiwhqkj5aTP25/wyZOXE71oHCghpzwGsIKZ+UNvDO9jre3l7Pmj3BzjdqIz1JOHhQEkW5yYzNSmJMlp8x7jUv8A8+y/KPaqjHV7od2bEN3/at+Lesx7dlPb7K8pizCo0eT6hoJqEpMwlNnmZdIhtjjDFRXFGU0SqoeX7zPn4yK5th6dH7dw7NPIpwfgG+shIAJNhA8nuvETz1wrgfb6KzoMYckLZWN/Ln7fW8vb2O97+upyaGh/kzkoTDh6Zw+JAUDs1L5pC8ZMZk+WPvArknBBuQmiqkcg9SUeZN5bvxVZYhpTvwlWxDykq7/SxMqHCMF8BMmUmoaLr9UmSMMcZ00fxRqYzL8rO52huPIRiG363by3/O6mBQTX8SwePPIvDsQ81JyW+/RHD+uda9cyfs6pgDQm2j8uFOL4h5Z3s9G9ygWF2RnSIcPSzAkcMCHFXgBTKRTcV6hCrU7fPGc6muJKf4M5J2bPAClpoqZG81srcKaqqQvVVIjfde6ut69DDCw0cTmjKTxikzCRdNt4fvjTHGmG7y+4Srp2Zy/dKWDgMeLq7huumZHQ5o7XXv/BjS4NXy+PaU4l/+IaHDj4n7MScyC2rMgLQ3GOZvuxr4qKSBj3bW8/ddDV0euFKAbwxO5rjCVE4oDDBrSEr3gpj6uuYgxRuAsgKprmgzX4FUl3vzwZZmb+Ni31vMNCXgPRMzbjLhcQcTmjIDzc3vhT0bY4wxB4YLJqRzx/IqKhq8FhPl9cofN+7jiqLM6BtlZtN45Ikkv7ewOSll8Z+otaCmQxbUmIQXDCvrKxpZVdbAyrIgy3Z5r7EMD1OQ5msOYo4dESAvNUp718ZgRDOvXfjKd3vzkUFKlQtSergWpbtUBM0vIDxsFOGCQsKjJxAeV0S4cIxVZRtjjDFxlJHs4/tFGdyzqmVA7QfW1nD55IwOm60HT/xOq6DGX7wS3xcbCB/U0RCeBzb7RmMSSm2j8ll5kJVlQS+I2RPks/Ig9V2shWmSJDCnIIUTClM5fmQqhwxKQlS9QGVLMVKyHV/5LqS8DKnYjZR7k6+6Ij4n1k0qPsjMIpw9CM0djA7KR3Py0UGDCQ8ajA4bSXhoIaQE+vpQjTHGmAPSVVMyWbC6mkb1gphNVSEWbavj5NFpUbcJjxpH45SZJK37tDkt8OQvqb3xXutlNAoLaky/VR0Ms7rMC2BWljWwak+Q4opGQt0c/3FUpp8TC1M5sUA4RkvILlmLb91mfO986fUOtuvrVk3Aepsmp6A5g9CsXGr8KaQXjEAzs9GMLDQzGzKy0IxsL82lk5puf9yMMcaYfmx4up95Q0K8Vtrytfv+tTUdBjUAwXnntApq/J+vIundV2k87sy4HWsis6DG9AuqyobKxuZnYJbvDrKxqusP87dnfLafb+Upp4a/5PDqzQz+egP+5ZuQndtiGo+lu9SfhGblotm5aPYgNCvHe21+33oZgTRwVdGbN2xg4kSrYjbGGGMGggtGBFsFNR/ubGDF7gZmDI4+bk1oxpE0TptN0qpPmtMCzzxEaMaRaN7QuB5vIrKgxvSZkn0hXv2ilvd31PNRSQO767ofaAxP9zEtL5njfaXMrfycol3FZCwvxrd9K6I9F8CoiBeE5A5GB3nNvcK5g10NS1PAkotm5UJ6ZnOQYowxxpgD1+RM5ejhAd7f0TJuzQNra/jNMXnRNxKh/rLr8N90GVJX6yXV7SPw+3up+9c77TtGGxbUmF5VXh/m1S9qeWFzLUt21hPuRlOygzL9TM9PZmauj6PrtnLIrvXkbF2Df8kapLqy8ww6oJnZ3sP0QwvR/ALv2ZSIAEZz8uzhemOMMcbE7Jqpma2Cmj9tqeXGmY2MzY7+vULzC2g4+yoCTy5oTkta8RFJf/0LjbOPi+vxJhr7dmbiTlVZsrOBX6+t4e3tdQS7WHEiwMScJKbnJzMtP5kZOcKs6s3kbFqNf9kK/BtWd7uHsXB+AeGRYwmPHOe9Dh/lPVCfkdWt/IwxxhhjOnLiyAATc5Kax8prVLjmg3IWnjy4457Qjj+LpKV/xr9xTXNayhMLaJw6ywbFjmBBjYmbsCoLv6jjvtXVLNvd+QP4WcnCEUNTOKogwJEFKUzLgextxfjXr8T/yQr8G9YiDbEHMeGCkYTGFREeV0RozCTCI8d5TcOMMcYYY3qJT4TrpmXxwyXlzWkflTTwwNoarj2kgx9VfT7qrvgR6T++Emn0vk/5qisIPPUr6n9wU7wPO2FYUGN6XDCsPL1xHwvW1DT/GhHNoXnJfGdsGseOCHBIlpKyZT3+4pX431yBf+Pa5tF0u0pT0wlNmEpo0qGExx9MaOxkq30xxhhjTL9w/vg0Xti8j7e3t3y/uX15FSeMTKUoNznqdjriIBrOuJjAnx5pTkv+6C0a5xxHaPqcuB5zorCgxvSYsCovbqnljuVVbKmOPnDMxJwkvjs2je+O9DO5bAP+dSvwv77SC2KCDbHtM2cQockzCE861AtkRo0DX5SBM40xxhhj+pCIsOCfBnHkSyVUNngPFteH4H+9X87i04aQ7OugGdqpF5D013fxf7W5OS31V7dS93/uJDR1VtyPvb+zoMb0iL9sr+PWZVWsLIvezOzM4cJNWds4pGQtSW+uxLfps5jHhQnn5hOaPJ3QlBmEJk9Hh4+23j+MMcYYkzBGZPi5e04uV73f0gxtRVmQn6+s5oaZ2dE3TEqm/op/J+22q5t7dpX6OlLvuYG6a35C6Btz433o/ZoFNeYf8unuBm79exXv7di/mVhaqJ65VRv4Z9nICVXryfxgfXNb0K4K5w72ApiiGYSKpqMFIy2IMcYYY0xCO3tcGgu/rOXlrS3PCt+9spp5o1KZ2cHYNeFxRTSccyWBZ3/TnCaNQVJ/eQv1V95A4z+dFNfj7s8sqDHdsrEyyB3Lq3lpa21zWlbjPo6q3MDcymKOrlzH7OrNJIVjG0AznDfEBTAuiBlaaEGMMcYYYwYUEeGeI3P5uKSU0lqv1qVR4Qfvl/PayYMZkha9KX3w1AvBn0Tgjw+05BcOk/qb/6K+di/BE74d9+PvjyyoMTHZuS/EXSuqePzzfeTXV/KdivXMrSxmbmUx02u+wE9sA8+E8wuaA5hQ0Qx0yHALYowxxhgz4OWn+rnvqFwueGdPc9rnlY0c80opjx2bxxFDA1G3Dc4/F03LIPDo3Yi2fPcKPHEfUlZKwxnfg7SMuB5/f2NBjemStWUNLPz7ZkpXrGDOnvVcX1nM5NodMecTHjxs/yDGGGOMMeYAdPLoNC6amM5TG/Y1p329L8wpr+/mjiNy+MGUDCTKj72Nx5wKqekEHroTCbW0jEl5/Y8kv/8aDadeSPD4syCQGvfz6A8sqDH7U0X2lNK4cT1bVn5G3ab1jC/bzO3B6pizCg8Z0fxQf6hoOjp4WBwO2BhjjDEmMd01O4cvqhv5YGdLD7CNCjd8UsknJQ0smJtLVrKv3W0bZx+LpqaTev8trYbBkJoqAs88SPKbzxA8/WKC3zoNkqM/qzMQ9FpQIyLzgfsAP/A7Vf3vNssDwOPALKAMOE9Vt7plNwJXACHgX1R1UUd5ishY4GkgD1gOXKyqsfUVfCBoqEf2lOLbsY2hq5cTeO8FGrZ9iX/nl6TW1QAwI8YswyMOIjRpGqFJh3o1MflDe/64jTHGGGMGiMxkHy/NG8xty6pYsKam1bIXt9bycUk9Z45J44wxacwZmoK/TbfPoemzqb3+f0j95S34qitaLfNVlhN4cgEpLz7qtZSZMpPQlJmEC8cMuOb+vRLUiIgf+BVwIvAV8DcReUVVP4tY7QqgXFUniMj5wF3AeSJyMHA+MBUYAbwtIpPcNtHyvAv4hao+LSIPurx/Hf8z7UPhMDTUIfV1UFeL1O5FaqqQmkqkpgpqqpDqCnx7StHdpUhZKUl7K5s3L3Sv0Yd92p/6fIQPmkRosgtiJh0KWbk9elrGGGOMMQNdkk+47fAcDh+awjVLyqkKtjwns7M2zEPr9vLQur0MTfNx2ug0ZhekMCLdT2GGn+HpftImT2Pfz54kZdFzJL/5HFK3r1X+sreapGVLSFq2BIBw9iDCEw8hPHgYmjeUcP4QNK8AzRuCpmdASmrCBT29VVNzBLBRVTcDiMjTwJlAZFBzJnCrm38euF+8RoRnAk+raj2wRUQ2uvxoL08RWQccB1zo1vm9y7dfBjVJ7y4k+YM38Z6vV1B1r0TMN70PQ2Oj124yFPTmG4NejUvD/l0q97Raf4DKUUXkTJsORdMIjT8YUtPjvl9jjDHGmAPB6QelcXBuMhf/pYzPyvfvQba0NswjxXt5pHhvq/S8gI/sFCHgO4Uhxx/N5Zte5byNb5Da2P73Q19VOT4X4LRHRSCQigZS0cHDqL2lX36NbkVUY+utqls7ETkbmK+qV7r3FwOzVfXaiHXWuHW+cu83AbPxApKlqvqkS38YeMNttl+eEetPcOmjgDdU9ZCmfVVWVsb/pI0xxhhjjDE9LicnZ79qpPafOup57dVftQ0soq3TU+nGGGOMMcaYAai3gpqvgFER70cCX0dbR0SSgBxgTwfbRkvfDeS6PKLtyxhjjDHGGDNA9NYzNX8DJrpeybbjPfh/YZt1XgEuBT4Gzgb+rKoqIq8AfxCRe/A6CpgI/BWvRma/PN02f3F5PO3yfDlyR+1VWRljjDHGGGMSU68ENaraKCLXAovwul9+RFXXishtwN9V9RXgYeAJ1xHAHrwgBbfes3idCjQC16hqCKC9PN0u/wN4WkTuAD51eRtjjDHGGGMGoN5qfoaqvq6qk1R1vKre6dJucQENqlqnqueo6gRVPaKpVzO37E633WRVfaOjPF36ZpfHBJdn/LsG68dE5HYRWSUiK0TkLREZ4dJFRBaIyEa3/BsR21wqIhvcdGlE+iwRWe22WSDRhrk1UYnI/4jIenfNXxSRXJc+RkRqXTmtcN2RN23T7nUXkTwRWezKabGIDOqr80pk0crELbvRXfdiEZkXkT7fpW0UkRsi0seKyCeuTJ4RkYE92lkciMg5IrJWRMIiclhEut0jfSBaebhldn/0MRG5VUS2R9wXp0Qsi6l8TM+za92LVNWmAT4B2RHz/wI86OZPwetJToA5wCcuPQ/Y7F4HuflBbtlfgSPdNm8AJ/f1+SXaBJwEJLn5u4C73PwYYE2Ubdq97sDPgBvc/A1NednUY2VyMLASCABjgU14NcN+Nz8OSHHrHOy2eRY4380/CPywr88v0SZgCjAZeBc4LCLd7pH+VR52f/SDCa/X1+vbSY+5fGzq8bKxa92LU6/V1Ji+o6pVEW8zaOkN7kzgcfUsxetgYTgwD1isqntUtRxYDMx3y7JV9WP17tbHgbN670wGBlV9S1WbOp9fiteZRVSdXPcz8cZiwr1aeXRDB2XSPE6Wqm4BmsbJah57S1Ub8J7fO9PVDhyHN9YWWJl0i6quU9Xirq5v90h8dVAedn/0bzGVTx8e50Bm17oXWVBzgBCRO0VkG3ARcItLLgS2Raz2lUvrKP2rdtJN932flnGXAMaKyKci8p6IfNOldXTdC1R1B4B7HRrvAz4ARJZJrPdIPlARESDZPdLz7B7pP+z+6D+udc1nH4loYhlr+ZieZ9e6F/VW72cmzkTkbWBYO4tuVtWXVfVm4GYRuRG4FvgJNgZQ3HRWHm6dm/E6v3jKLdsBjFbVMhGZBbwkIlOx694julkm0a59ez8I2T0Sg66URzvsHomTbpaH3R+9pKPyAX4N3I53LW8Hfo7340ys5WN6nn3me5EFNQOEqp7QxVX/ALyGF9R0NAbQt9qkv+vSR7azvmmjs/IQr/OF04DjXXMZ1OvQot7NLxORTcAkOr7uJSIyXFV3uCY4pT17JgNHd8qEjsfY6nCcLPdrtN0jUcTwNytyG7tH4qQ75YHdH72mq+UjIr8FFrq3sZaP6XldGafR9BBrfnYAEJGJEW/PANa7+VeAS8QzB6h0zTMWASeJyCBXjX0SsMgtqxaROa5t9CW0GQPIdE5E5uN1O36Gqu6LSB8iIn43Pw5vTKbNnVz3pvGdoJ0xmUzXRCsTvOt7vogExBsTq2mcrOaxt1zvTecDr7hgqGmcLLAy6VF2j/Q7dn/0Ay5Yb/JtYI2bj6l8evOYDyB2rXtTX/dUYFP8J+AFvD9yq4BXgUKXLsCv8HrmWE3rXm2+j/dQ4Ubg8oj0w1xem4D7Aenr80u0yV3TbcAKNzX1RvddYC1e7yjLgdM7u+54bdTfATa417y+Pr9EnKKViVt2s7vuxUT09ofXe+DnbtnNEenj8L44bASeAwJ9fX6JNuF9MfsKr1amBO9HFbtH+ll5uGV2f/R9+Tzh/oevwvvCPLy75WNTXMrHrnUvTU1/9I0xxhhjjDEmIVnzM2OMMcYYY0xCs6DGGGOMMcYYk9AsqDHGGGOMMcYkNAtqjDHGGGOMMQnNghpjjDHGGGNMQrOgxhhjjDHGGJPQLKgxxpg+JCJbRaQ7o7kbY4wxxrGgxhhjTLeIyBgRURGpcVOJiCwUkROjrP+uiJSLSCAi7SoRWdcmLV9ESkVkvnt/k4hscfv4SkSe6eLxzROR90WkWkR2ich7InKGW3aZiIRcnlUislJETuvg3Jqm89zyx0SkweVdLSJrROSnIpITkcdlIvKBm4/MIywitRHvL4rtyhtjjGnLghpjjBkARMTfh7vPVdVMYDqwGHhRRC6LXEFExgDfBBQ4oyldVX+LN1r9LRGr3wu8rqpvisilwMXACW4fhwHvdHZAInI23oj1jwMjgQK3j9MjVvvY5ZkLPAA8LSK57Z1bxBQZUP1MVbOAIcDlwBzgQxHJaHs8kXkAXwKnR6Q91dn5GGOM6ZgFNcYY0/cOF5HPXC3GoyKSGvkrfxNXczDBzT8mIr8WkddFZC9wrIgERORuEfnS1Zo8KCJpbv1BrhZll9vPQhEZGZF3q2ZwInKriDwZy0mo6k5VvQ+4FbhLRCL/x1wCLAUeAy5ts+lVwNUiMkNETgKOB/6t6doAi1R1U8Q+ftPRcYiIAPcAt6vq71S1UlXDqvqeql7VznGHgSeADGBiLOfstq9T1b/hBWv5eAFOj3Dl8JyIPOlqhFaLyCQRudHVZm1z16xp/RwReVhEdojIdhG5oyngFZHxIvJnESkTkd0i8lRkEOc+A9eLyCoRqRSRZ0QktafOxRhj4smCGmOM6XsXAfOA8cAk4D+7uN2FwJ1AFvABcJfbfgYwASikpQbEBzwKHASMBmqB+3vm8PfzJ2AoMDki7RLgKTfNE5GCpgWqutUd5yPAQ8DVqlruFi8FLhGRH4nIYV2skZoMjAKe78rBujwvB4LAF13Zpj2qWo1XU/XN7uYRxel4Qdcg4FNgEV55FgK34V2zJr8HGvHKfyZwEnClWybAT4ERwBS8a3Rrm32dC8wHxgLTgMt6+FyMMSYuLKgxxpi+d7+qblPVPXhBygVd3O5lVf3Q1TTU49V4/Juq7nFfsP8LOB9AVctU9QVV3eeW3Qkc0/OnAsDX7jUPQETm4gVTz6rqMmATXkAW6X68oGKFqr7UlKiqTwL/Gy/oew8oFZEbOtl/vnvd0Yj6e3MAAAO/SURBVMl6c0SkAqgD7ga+p6qlbdbZLSIVEdOUTvL8GnfePWiJqi5S1Ua8JnVDgP9W1SDwNDBGRHJdoHgy8K+qutedyy9o+QxsVNXFqlqvqrvwarPafgYWqOrX7rP4Kl6AbIwx/V5SXx+AMcYYtkXMf4H3S3qs2w0B0oFlXusrwPtlvqnpUTreF9z5eL/4A2SJiF9VQ9087mgK3ese93op8Jaq7nbv/+DSftG0gaqqiKzDe76mFffMyVMikgyc5eY/VdVFUfZf5l6HA1s6OM6lqjpXRDKBh/FqWJ5ts85gF0x0VSEt591TSiLma4HdEWVW614z8T43ycCOiM+AD/c5EZGhwAK888xyy5pqxJrsjJjfR9c/i8YY06espsYYY/reqIj50Xi/9u/FC1IAEJFh7WynEfO78b7gTlXVXDfluAfTAf4vXrOs2aqaDRzdlLV7bbU/oL39ddW3gVKg2D3Tcy5wjIjsFJGdeM/LTBeR6bFkqqpBVX0OWAUc0sGqxXhf5L/bxXxrgKuBi0VkZizHFMkFRycAS7qbxz9oG16N3eCIz0C2qk51y3+K95mZ5j4D36Ol/I0xJqFZUGOMMX3vGhEZKSJ5wE3AM8BKYKp7eD6V/Z99aMU1Qfst8Av3izwiUigi89wqWXhBT4Xbz0/aZLECOF9EkkXkMODsWE9CRApE5FqX943umM4CQsDBeE2ZZuA9z7EE7zmbzvK8TEROFZEsEfGJyMnAVOCTaNuoqgLXAT8WkctFJNttO1dE2u1kQFXLgN/Ruhe2LnEdNMwCXsKr+Xg01jx6gqruAN4Cfh5xzuNFpKmJWRZQg/cZKAR+1BfHaYwx8WBBjTHG9L0/4H0Z3eymO1T1c7yHwN8GNuB1BNCZ/wA2AktFpMpt2/Sw/r1AGl6NzlLgzTbb/hivo4Jy4P+5Y+qqCtcD22rgFOAcVX3ELbsUeFRVv3Q9l+1U1Z14z9BcJCKdNYOuwgv0vgQqgJ8BP1TVDq+Hqj4PnAd8H6/mqwS4A3i5g83uBU4RkWltzi1yjJnrIpb9u4hU4zU3exxYBhylqns7Oad4ugRIAT7DK8vn8ZrhgVeu3wAqgdfwOnQwxpgBQbwftIwxxhhjjDEmMVlNjTHGGGOMMSahWVBjjDEmKhG5qE3zq6ZpbT84tvaOq0ZEenqcmLgSkTeinMdNfX1sxhiTKKz5mTHGGGOMMSahWU2NMcYYY4wxJqFZUGOMMcYYY4xJaBbUGGOMMcYYYxKaBTXGGGOMMcaYhGZBjTHGGGOMMSah/X9602qSixfzCAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "kde_target('bureau_DAYS_CREDIT_mean', train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "58ca4b2852b486fa10e1bf31c2a7e61e9de52dd7"
   },
   "source": [
    "The definition of this column is: \"How many days before current application did client apply for Credit Bureau credit\". My interpretation is this is the number of days that the previous loan was applied for before the application for a loan at Home Credit. Therefore, a larger negative number indicates the loan was further before the current loan application. We see an extremely weak positive relationship between the average of this variable and the target meaning that clients who applied for loans further in the past potentially are more likely to repay loans at Home Credit. With a correlation this weak though, it is just as likely to be noise as a signal. \n",
    "\n",
    "#### The Multiple Comparisons Problem\n",
    "\n",
    "When we have lots of variables, we expect some of them to be correlated just by pure chance, a [problem known as multiple comparisons](https://towardsdatascience.com/the-multiple-comparisons-problem-e5573e8b9578). We can make hundreds of features, and some will turn out to be corelated with the target simply because of random noise in the data. Then, when our model trains, it may overfit to these variables because it thinks they have a relationship with the target in the training set, but this does not necessarily generalize to the test set. There are many considerations that we have to take into account when making features! "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "645ab51cb64cd20f122d61e17ef8ad1ed78443cc"
   },
   "source": [
    "## Function for Numeric Aggregations\n",
    "\n",
    "Let's encapsulate all of the previous work into a function. This will allow us to compute aggregate stats for numeric columns across any dataframe. We will re-use this function when we want to apply the same operations for other dataframes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "_uuid": "8203852b47164d7800f69e572dda2cbe83c50976",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def agg_numeric(df, group_var, df_name):\n",
    "    \"\"\"Aggregates the numeric values in a dataframe. This can\n",
    "    be used to create features for each instance of the grouping variable.\n",
    "    \n",
    "    Args:\n",
    "        df (dataframe): the dataframe to calculate the statistics on\n",
    "        group_var (string): the variable by which to group df\n",
    "        df_name (string): the variable used to rename the columns\n",
    "        \n",
    "    Returns:\n",
    "        agg (dataframe): a dataframe with the statistics aggregated for \n",
    "        all numeric columns. Each instance of the grouping variable will have \n",
    "        the statistics (mean, min, max, sum; currently supported) calculated. \n",
    "        The columns are also renamed to keep track of features created.\n",
    "    \n",
    "    \"\"\"\n",
    "    \n",
    "    # First calculate counts\n",
    "    counts = pd.DataFrame(df.groupby(group_var, as_index = False)[df.columns[1]].count()).rename(columns = {df.columns[1]: '%s_counts' % df_name})\n",
    "    \n",
    "    # Group by the specified variable and calculate the statistics\n",
    "    agg = df.groupby(group_var).agg(['mean', 'max', 'min', 'sum']).reset_index()\n",
    "    \n",
    "    # Need to create new column names\n",
    "    columns = [group_var]\n",
    "    \n",
    "    # Iterate through the variables names\n",
    "    for var in agg.columns.levels[0]:\n",
    "        # Skip the grouping variable\n",
    "        if var != group_var:\n",
    "            # Iterate through the stat names\n",
    "            for stat in agg.columns.levels[1][:-1]:\n",
    "                # Make a new column name for the variable and stat\n",
    "                columns.append('%s_%s_%s' % (df_name, var, stat))\n",
    "              \n",
    "    #  Rename the columns\n",
    "    agg.columns = columns\n",
    "    \n",
    "    # Merge with the counts\n",
    "    agg = agg.merge(counts, on = group_var, how = 'left')\n",
    "    \n",
    "    return agg"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "_uuid": "2784423d6e900b66f108cb2e6a425d63eb28743e",
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th>bureau_DAYS_CREDIT_mean</th>\n",
       "      <th>bureau_DAYS_CREDIT_max</th>\n",
       "      <th>bureau_DAYS_CREDIT_min</th>\n",
       "      <th>bureau_DAYS_CREDIT_sum</th>\n",
       "      <th>bureau_CREDIT_DAY_OVERDUE_mean</th>\n",
       "      <th>bureau_CREDIT_DAY_OVERDUE_max</th>\n",
       "      <th>bureau_CREDIT_DAY_OVERDUE_min</th>\n",
       "      <th>bureau_CREDIT_DAY_OVERDUE_sum</th>\n",
       "      <th>bureau_DAYS_CREDIT_ENDDATE_mean</th>\n",
       "      <th>...</th>\n",
       "      <th>bureau_AMT_CREDIT_SUM_OVERDUE_sum</th>\n",
       "      <th>bureau_DAYS_CREDIT_UPDATE_mean</th>\n",
       "      <th>bureau_DAYS_CREDIT_UPDATE_max</th>\n",
       "      <th>bureau_DAYS_CREDIT_UPDATE_min</th>\n",
       "      <th>bureau_DAYS_CREDIT_UPDATE_sum</th>\n",
       "      <th>bureau_AMT_ANNUITY_mean</th>\n",
       "      <th>bureau_AMT_ANNUITY_max</th>\n",
       "      <th>bureau_AMT_ANNUITY_min</th>\n",
       "      <th>bureau_AMT_ANNUITY_sum</th>\n",
       "      <th>bureau_counts</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>100001</td>\n",
       "      <td>-735.000000</td>\n",
       "      <td>-49</td>\n",
       "      <td>-1572</td>\n",
       "      <td>-5145</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>82.428571</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-93.142857</td>\n",
       "      <td>-6</td>\n",
       "      <td>-155</td>\n",
       "      <td>-652</td>\n",
       "      <td>3545.357143</td>\n",
       "      <td>10822.5</td>\n",
       "      <td>0.0</td>\n",
       "      <td>24817.5</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>100002</td>\n",
       "      <td>-874.000000</td>\n",
       "      <td>-103</td>\n",
       "      <td>-1437</td>\n",
       "      <td>-6992</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-349.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-499.875000</td>\n",
       "      <td>-7</td>\n",
       "      <td>-1185</td>\n",
       "      <td>-3999</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>100003</td>\n",
       "      <td>-1400.750000</td>\n",
       "      <td>-606</td>\n",
       "      <td>-2586</td>\n",
       "      <td>-5603</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-544.500000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-816.000000</td>\n",
       "      <td>-43</td>\n",
       "      <td>-2131</td>\n",
       "      <td>-3264</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>100004</td>\n",
       "      <td>-867.000000</td>\n",
       "      <td>-408</td>\n",
       "      <td>-1326</td>\n",
       "      <td>-1734</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-488.500000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-532.000000</td>\n",
       "      <td>-382</td>\n",
       "      <td>-682</td>\n",
       "      <td>-1064</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>100005</td>\n",
       "      <td>-190.666667</td>\n",
       "      <td>-62</td>\n",
       "      <td>-373</td>\n",
       "      <td>-572</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>439.333333</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-54.333333</td>\n",
       "      <td>-11</td>\n",
       "      <td>-121</td>\n",
       "      <td>-163</td>\n",
       "      <td>1420.500000</td>\n",
       "      <td>4261.5</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4261.5</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 50 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_CURR  bureau_DAYS_CREDIT_mean  bureau_DAYS_CREDIT_max  \\\n",
       "0      100001              -735.000000                     -49   \n",
       "1      100002              -874.000000                    -103   \n",
       "2      100003             -1400.750000                    -606   \n",
       "3      100004              -867.000000                    -408   \n",
       "4      100005              -190.666667                     -62   \n",
       "\n",
       "   bureau_DAYS_CREDIT_min  bureau_DAYS_CREDIT_sum  \\\n",
       "0                   -1572                   -5145   \n",
       "1                   -1437                   -6992   \n",
       "2                   -2586                   -5603   \n",
       "3                   -1326                   -1734   \n",
       "4                    -373                    -572   \n",
       "\n",
       "   bureau_CREDIT_DAY_OVERDUE_mean  bureau_CREDIT_DAY_OVERDUE_max  \\\n",
       "0                             0.0                              0   \n",
       "1                             0.0                              0   \n",
       "2                             0.0                              0   \n",
       "3                             0.0                              0   \n",
       "4                             0.0                              0   \n",
       "\n",
       "   bureau_CREDIT_DAY_OVERDUE_min  bureau_CREDIT_DAY_OVERDUE_sum  \\\n",
       "0                              0                              0   \n",
       "1                              0                              0   \n",
       "2                              0                              0   \n",
       "3                              0                              0   \n",
       "4                              0                              0   \n",
       "\n",
       "   bureau_DAYS_CREDIT_ENDDATE_mean      ...        \\\n",
       "0                        82.428571      ...         \n",
       "1                      -349.000000      ...         \n",
       "2                      -544.500000      ...         \n",
       "3                      -488.500000      ...         \n",
       "4                       439.333333      ...         \n",
       "\n",
       "   bureau_AMT_CREDIT_SUM_OVERDUE_sum  bureau_DAYS_CREDIT_UPDATE_mean  \\\n",
       "0                                0.0                      -93.142857   \n",
       "1                                0.0                     -499.875000   \n",
       "2                                0.0                     -816.000000   \n",
       "3                                0.0                     -532.000000   \n",
       "4                                0.0                      -54.333333   \n",
       "\n",
       "   bureau_DAYS_CREDIT_UPDATE_max  bureau_DAYS_CREDIT_UPDATE_min  \\\n",
       "0                             -6                           -155   \n",
       "1                             -7                          -1185   \n",
       "2                            -43                          -2131   \n",
       "3                           -382                           -682   \n",
       "4                            -11                           -121   \n",
       "\n",
       "   bureau_DAYS_CREDIT_UPDATE_sum  bureau_AMT_ANNUITY_mean  \\\n",
       "0                           -652              3545.357143   \n",
       "1                          -3999                 0.000000   \n",
       "2                          -3264                      NaN   \n",
       "3                          -1064                      NaN   \n",
       "4                           -163              1420.500000   \n",
       "\n",
       "   bureau_AMT_ANNUITY_max  bureau_AMT_ANNUITY_min  bureau_AMT_ANNUITY_sum  \\\n",
       "0                 10822.5                     0.0                 24817.5   \n",
       "1                     0.0                     0.0                     0.0   \n",
       "2                     NaN                     NaN                     0.0   \n",
       "3                     NaN                     NaN                     0.0   \n",
       "4                  4261.5                     0.0                  4261.5   \n",
       "\n",
       "   bureau_counts  \n",
       "0              7  \n",
       "1              8  \n",
       "2              4  \n",
       "3              2  \n",
       "4              3  \n",
       "\n",
       "[5 rows x 50 columns]"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bureau_agg_new = agg_numeric(bureau.drop(columns = ['SK_ID_BUREAU']), group_var = 'SK_ID_CURR', df_name = 'bureau')\n",
    "bureau_agg_new.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "171660155dd332d9df85f61fdd7004a9f2593e59"
   },
   "source": [
    "To make sure the function worked as intended, we should compare with the aggregated dataframe we constructed by hand. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "_uuid": "57dab1b1b23d0bd0167c9367688b56b43066a721",
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th>bureau_DAYS_CREDIT_mean</th>\n",
       "      <th>bureau_DAYS_CREDIT_max</th>\n",
       "      <th>bureau_DAYS_CREDIT_min</th>\n",
       "      <th>bureau_DAYS_CREDIT_sum</th>\n",
       "      <th>bureau_CREDIT_DAY_OVERDUE_mean</th>\n",
       "      <th>bureau_CREDIT_DAY_OVERDUE_max</th>\n",
       "      <th>bureau_CREDIT_DAY_OVERDUE_min</th>\n",
       "      <th>bureau_CREDIT_DAY_OVERDUE_sum</th>\n",
       "      <th>bureau_DAYS_CREDIT_ENDDATE_mean</th>\n",
       "      <th>...</th>\n",
       "      <th>bureau_AMT_CREDIT_SUM_OVERDUE_sum</th>\n",
       "      <th>bureau_DAYS_CREDIT_UPDATE_mean</th>\n",
       "      <th>bureau_DAYS_CREDIT_UPDATE_max</th>\n",
       "      <th>bureau_DAYS_CREDIT_UPDATE_min</th>\n",
       "      <th>bureau_DAYS_CREDIT_UPDATE_sum</th>\n",
       "      <th>bureau_AMT_ANNUITY_mean</th>\n",
       "      <th>bureau_AMT_ANNUITY_max</th>\n",
       "      <th>bureau_AMT_ANNUITY_min</th>\n",
       "      <th>bureau_AMT_ANNUITY_sum</th>\n",
       "      <th>bureau_counts</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>100001</td>\n",
       "      <td>-735.000000</td>\n",
       "      <td>-49</td>\n",
       "      <td>-1572</td>\n",
       "      <td>-5145</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>82.428571</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-93.142857</td>\n",
       "      <td>-6</td>\n",
       "      <td>-155</td>\n",
       "      <td>-652</td>\n",
       "      <td>3545.357143</td>\n",
       "      <td>10822.5</td>\n",
       "      <td>0.0</td>\n",
       "      <td>24817.5</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>100002</td>\n",
       "      <td>-874.000000</td>\n",
       "      <td>-103</td>\n",
       "      <td>-1437</td>\n",
       "      <td>-6992</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-349.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-499.875000</td>\n",
       "      <td>-7</td>\n",
       "      <td>-1185</td>\n",
       "      <td>-3999</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>100003</td>\n",
       "      <td>-1400.750000</td>\n",
       "      <td>-606</td>\n",
       "      <td>-2586</td>\n",
       "      <td>-5603</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-544.500000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-816.000000</td>\n",
       "      <td>-43</td>\n",
       "      <td>-2131</td>\n",
       "      <td>-3264</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>100004</td>\n",
       "      <td>-867.000000</td>\n",
       "      <td>-408</td>\n",
       "      <td>-1326</td>\n",
       "      <td>-1734</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-488.500000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-532.000000</td>\n",
       "      <td>-382</td>\n",
       "      <td>-682</td>\n",
       "      <td>-1064</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>100005</td>\n",
       "      <td>-190.666667</td>\n",
       "      <td>-62</td>\n",
       "      <td>-373</td>\n",
       "      <td>-572</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>439.333333</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-54.333333</td>\n",
       "      <td>-11</td>\n",
       "      <td>-121</td>\n",
       "      <td>-163</td>\n",
       "      <td>1420.500000</td>\n",
       "      <td>4261.5</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4261.5</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 50 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_CURR  bureau_DAYS_CREDIT_mean  bureau_DAYS_CREDIT_max  \\\n",
       "0      100001              -735.000000                     -49   \n",
       "1      100002              -874.000000                    -103   \n",
       "2      100003             -1400.750000                    -606   \n",
       "3      100004              -867.000000                    -408   \n",
       "4      100005              -190.666667                     -62   \n",
       "\n",
       "   bureau_DAYS_CREDIT_min  bureau_DAYS_CREDIT_sum  \\\n",
       "0                   -1572                   -5145   \n",
       "1                   -1437                   -6992   \n",
       "2                   -2586                   -5603   \n",
       "3                   -1326                   -1734   \n",
       "4                    -373                    -572   \n",
       "\n",
       "   bureau_CREDIT_DAY_OVERDUE_mean  bureau_CREDIT_DAY_OVERDUE_max  \\\n",
       "0                             0.0                              0   \n",
       "1                             0.0                              0   \n",
       "2                             0.0                              0   \n",
       "3                             0.0                              0   \n",
       "4                             0.0                              0   \n",
       "\n",
       "   bureau_CREDIT_DAY_OVERDUE_min  bureau_CREDIT_DAY_OVERDUE_sum  \\\n",
       "0                              0                              0   \n",
       "1                              0                              0   \n",
       "2                              0                              0   \n",
       "3                              0                              0   \n",
       "4                              0                              0   \n",
       "\n",
       "   bureau_DAYS_CREDIT_ENDDATE_mean      ...        \\\n",
       "0                        82.428571      ...         \n",
       "1                      -349.000000      ...         \n",
       "2                      -544.500000      ...         \n",
       "3                      -488.500000      ...         \n",
       "4                       439.333333      ...         \n",
       "\n",
       "   bureau_AMT_CREDIT_SUM_OVERDUE_sum  bureau_DAYS_CREDIT_UPDATE_mean  \\\n",
       "0                                0.0                      -93.142857   \n",
       "1                                0.0                     -499.875000   \n",
       "2                                0.0                     -816.000000   \n",
       "3                                0.0                     -532.000000   \n",
       "4                                0.0                      -54.333333   \n",
       "\n",
       "   bureau_DAYS_CREDIT_UPDATE_max  bureau_DAYS_CREDIT_UPDATE_min  \\\n",
       "0                             -6                           -155   \n",
       "1                             -7                          -1185   \n",
       "2                            -43                          -2131   \n",
       "3                           -382                           -682   \n",
       "4                            -11                           -121   \n",
       "\n",
       "   bureau_DAYS_CREDIT_UPDATE_sum  bureau_AMT_ANNUITY_mean  \\\n",
       "0                           -652              3545.357143   \n",
       "1                          -3999                 0.000000   \n",
       "2                          -3264                      NaN   \n",
       "3                          -1064                      NaN   \n",
       "4                           -163              1420.500000   \n",
       "\n",
       "   bureau_AMT_ANNUITY_max  bureau_AMT_ANNUITY_min  bureau_AMT_ANNUITY_sum  \\\n",
       "0                 10822.5                     0.0                 24817.5   \n",
       "1                     0.0                     0.0                     0.0   \n",
       "2                     NaN                     NaN                     0.0   \n",
       "3                     NaN                     NaN                     0.0   \n",
       "4                  4261.5                     0.0                  4261.5   \n",
       "\n",
       "   bureau_counts  \n",
       "0              7  \n",
       "1              8  \n",
       "2              4  \n",
       "3              2  \n",
       "4              3  \n",
       "\n",
       "[5 rows x 50 columns]"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bureau_agg['bureau_counts'] = bureau_agg_new['bureau_counts']\n",
    "bureau_agg.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "af5a3d2ebc6850707c0bb021bc08df50b98d008b"
   },
   "source": [
    "If we go through and inspect the values, we do find that they are equivalent. We will be able to reuse this function for calculating numeric stats for other dataframes. Using functions allows for consistent results and decreases the amount of work we have to do in the future! \n",
    "\n",
    "### Correlation Function\n",
    "\n",
    "Before we move on, we can also make the code to calculate correlations with the target into a function."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "_uuid": "f55cdb9e28c9fcdb16b7922ce90c99ddda590e67",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Function to calculate correlations with the target for a dataframe\n",
    "def target_corrs(df):\n",
    "\n",
    "    # List of correlations\n",
    "    corrs = []\n",
    "\n",
    "    # Iterate through the columns \n",
    "    for col in df.columns:\n",
    "        print(col)\n",
    "        # Skip the target column\n",
    "        if col != 'TARGET':\n",
    "            # Calculate correlation with the target\n",
    "            corr = df['TARGET'].corr(df[col])\n",
    "\n",
    "            # Append the list as a tuple\n",
    "            corrs.append((col, corr))\n",
    "            \n",
    "    # Sort by absolute magnitude of correlations\n",
    "    corrs = sorted(corrs, key = lambda x: abs(x[1]), reverse = True)\n",
    "    \n",
    "    return corrs"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "a86ad170924ebf66122d24083679d2a061a080ff"
   },
   "source": [
    "## Categorical Variables\n",
    "\n",
    "Now we move from the numeric columns to the categorical columns. These are discrete string variables, so we cannot just calculate statistics such as mean \n",
    "and max which only work with numeric variables. Instead, we will rely on calculating value counts of each category within each categorical variable. As an example, if we have the following dataframe:\n",
    "\n",
    "| SK_ID_CURR | Loan type |\n",
    "|------------|-----------|\n",
    "| 1          | home      |\n",
    "| 1          | home      |\n",
    "| 1          | home      |\n",
    "| 1          | credit    |\n",
    "| 2          | credit    |\n",
    "| 3          | credit    |\n",
    "| 3          | cash      |\n",
    "| 3          | cash      |\n",
    "| 4          | credit    |\n",
    "| 4          | home      |\n",
    "| 4          | home      |\n",
    "\n",
    "we will use this information counting the number of loans in each category for each client. \n",
    "\n",
    "| SK_ID_CURR | credit count | cash count | home count | total count |\n",
    "|------------|--------------|------------|------------|-------------|\n",
    "| 1          | 1            | 0          | 3          | 4           |\n",
    "| 2          | 1            | 0          | 0          | 1           |\n",
    "| 3          | 1            | 2          | 0          | 3           |\n",
    "| 4          | 1            | 0          | 2          | 3           |\n",
    "\n",
    "\n",
    "Then we can normalize these value counts by the total number of occurences of that categorical variable for that observation (meaning that the normalized counts must sum to 1.0 for each observation).\n",
    "\n",
    "| SK_ID_CURR | credit count | cash count | home count | total count | credit count norm | cash count norm | home count norm |\n",
    "|------------|--------------|------------|------------|-------------|-------------------|-----------------|-----------------|\n",
    "| 1          | 1            | 0          | 3          | 4           | 0.25              | 0               | 0.75            |\n",
    "| 2          | 1            | 0          | 0          | 1           | 1.00              | 0               | 0               |\n",
    "| 3          | 1            | 2          | 0          | 3           | 0.33              | 0.66            | 0               |\n",
    "| 4          | 1            | 0          | 2          | 3           | 0.33              | 0               | 0.66            |\n",
    "\n",
    "Hopefully, encoding the categorical variables this way will allow us to capture the information they contain. If anyone has a better idea for this process, please let me know in the comments!\n",
    "We will now go through this process step-by-step. At the end, we will wrap up all the code into one function to be re-used for many dataframes."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "3a8c1a8a74741d727a17ec0e83c237abb931e89e"
   },
   "source": [
    "First we one-hot encode a dataframe with only the categorical columns (`dtype == 'object'`)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>CREDIT_ACTIVE_Active</th>\n",
       "      <th>CREDIT_ACTIVE_Bad debt</th>\n",
       "      <th>CREDIT_ACTIVE_Closed</th>\n",
       "      <th>CREDIT_ACTIVE_Sold</th>\n",
       "      <th>CREDIT_CURRENCY_currency 1</th>\n",
       "      <th>CREDIT_CURRENCY_currency 2</th>\n",
       "      <th>CREDIT_CURRENCY_currency 3</th>\n",
       "      <th>CREDIT_CURRENCY_currency 4</th>\n",
       "      <th>CREDIT_TYPE_Another type of loan</th>\n",
       "      <th>CREDIT_TYPE_Car loan</th>\n",
       "      <th>...</th>\n",
       "      <th>CREDIT_TYPE_Loan for business development</th>\n",
       "      <th>CREDIT_TYPE_Loan for purchase of shares (margin lending)</th>\n",
       "      <th>CREDIT_TYPE_Loan for the purchase of equipment</th>\n",
       "      <th>CREDIT_TYPE_Loan for working capital replenishment</th>\n",
       "      <th>CREDIT_TYPE_Microloan</th>\n",
       "      <th>CREDIT_TYPE_Mobile operator loan</th>\n",
       "      <th>CREDIT_TYPE_Mortgage</th>\n",
       "      <th>CREDIT_TYPE_Real estate loan</th>\n",
       "      <th>CREDIT_TYPE_Unknown type of loan</th>\n",
       "      <th>SK_ID_CURR</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>215354</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>215354</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>215354</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>215354</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>215354</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 24 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   CREDIT_ACTIVE_Active  CREDIT_ACTIVE_Bad debt  CREDIT_ACTIVE_Closed  \\\n",
       "0                     0                       0                     1   \n",
       "1                     1                       0                     0   \n",
       "2                     1                       0                     0   \n",
       "3                     1                       0                     0   \n",
       "4                     1                       0                     0   \n",
       "\n",
       "   CREDIT_ACTIVE_Sold  CREDIT_CURRENCY_currency 1  CREDIT_CURRENCY_currency 2  \\\n",
       "0                   0                           1                           0   \n",
       "1                   0                           1                           0   \n",
       "2                   0                           1                           0   \n",
       "3                   0                           1                           0   \n",
       "4                   0                           1                           0   \n",
       "\n",
       "   CREDIT_CURRENCY_currency 3  CREDIT_CURRENCY_currency 4  \\\n",
       "0                           0                           0   \n",
       "1                           0                           0   \n",
       "2                           0                           0   \n",
       "3                           0                           0   \n",
       "4                           0                           0   \n",
       "\n",
       "   CREDIT_TYPE_Another type of loan  CREDIT_TYPE_Car loan     ...      \\\n",
       "0                                 0                     0     ...       \n",
       "1                                 0                     0     ...       \n",
       "2                                 0                     0     ...       \n",
       "3                                 0                     0     ...       \n",
       "4                                 0                     0     ...       \n",
       "\n",
       "   CREDIT_TYPE_Loan for business development  \\\n",
       "0                                          0   \n",
       "1                                          0   \n",
       "2                                          0   \n",
       "3                                          0   \n",
       "4                                          0   \n",
       "\n",
       "   CREDIT_TYPE_Loan for purchase of shares (margin lending)  \\\n",
       "0                                                  0          \n",
       "1                                                  0          \n",
       "2                                                  0          \n",
       "3                                                  0          \n",
       "4                                                  0          \n",
       "\n",
       "   CREDIT_TYPE_Loan for the purchase of equipment  \\\n",
       "0                                               0   \n",
       "1                                               0   \n",
       "2                                               0   \n",
       "3                                               0   \n",
       "4                                               0   \n",
       "\n",
       "   CREDIT_TYPE_Loan for working capital replenishment  CREDIT_TYPE_Microloan  \\\n",
       "0                                                  0                       0   \n",
       "1                                                  0                       0   \n",
       "2                                                  0                       0   \n",
       "3                                                  0                       0   \n",
       "4                                                  0                       0   \n",
       "\n",
       "   CREDIT_TYPE_Mobile operator loan  CREDIT_TYPE_Mortgage  \\\n",
       "0                                 0                     0   \n",
       "1                                 0                     0   \n",
       "2                                 0                     0   \n",
       "3                                 0                     0   \n",
       "4                                 0                     0   \n",
       "\n",
       "   CREDIT_TYPE_Real estate loan  CREDIT_TYPE_Unknown type of loan  SK_ID_CURR  \n",
       "0                             0                                 0      215354  \n",
       "1                             0                                 0      215354  \n",
       "2                             0                                 0      215354  \n",
       "3                             0                                 0      215354  \n",
       "4                             0                                 0      215354  \n",
       "\n",
       "[5 rows x 24 columns]"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "categorical = pd.get_dummies(bureau.select_dtypes('object'))\n",
    "categorical['SK_ID_CURR'] = bureau['SK_ID_CURR']\n",
    "categorical.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"2\" halign=\"left\">CREDIT_ACTIVE_Active</th>\n",
       "      <th colspan=\"2\" halign=\"left\">CREDIT_ACTIVE_Bad debt</th>\n",
       "      <th colspan=\"2\" halign=\"left\">CREDIT_ACTIVE_Closed</th>\n",
       "      <th colspan=\"2\" halign=\"left\">CREDIT_ACTIVE_Sold</th>\n",
       "      <th colspan=\"2\" halign=\"left\">CREDIT_CURRENCY_currency 1</th>\n",
       "      <th>...</th>\n",
       "      <th colspan=\"2\" halign=\"left\">CREDIT_TYPE_Microloan</th>\n",
       "      <th colspan=\"2\" halign=\"left\">CREDIT_TYPE_Mobile operator loan</th>\n",
       "      <th colspan=\"2\" halign=\"left\">CREDIT_TYPE_Mortgage</th>\n",
       "      <th colspan=\"2\" halign=\"left\">CREDIT_TYPE_Real estate loan</th>\n",
       "      <th colspan=\"2\" halign=\"left\">CREDIT_TYPE_Unknown type of loan</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>sum</th>\n",
       "      <th>mean</th>\n",
       "      <th>sum</th>\n",
       "      <th>mean</th>\n",
       "      <th>sum</th>\n",
       "      <th>mean</th>\n",
       "      <th>sum</th>\n",
       "      <th>mean</th>\n",
       "      <th>sum</th>\n",
       "      <th>mean</th>\n",
       "      <th>...</th>\n",
       "      <th>sum</th>\n",
       "      <th>mean</th>\n",
       "      <th>sum</th>\n",
       "      <th>mean</th>\n",
       "      <th>sum</th>\n",
       "      <th>mean</th>\n",
       "      <th>sum</th>\n",
       "      <th>mean</th>\n",
       "      <th>sum</th>\n",
       "      <th>mean</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>100001</th>\n",
       "      <td>3</td>\n",
       "      <td>0.428571</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4</td>\n",
       "      <td>0.571429</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>7</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100002</th>\n",
       "      <td>2</td>\n",
       "      <td>0.250000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6</td>\n",
       "      <td>0.750000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>8</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100003</th>\n",
       "      <td>1</td>\n",
       "      <td>0.250000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3</td>\n",
       "      <td>0.750000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100004</th>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100005</th>\n",
       "      <td>2</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 46 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "           CREDIT_ACTIVE_Active           CREDIT_ACTIVE_Bad debt       \\\n",
       "                            sum      mean                    sum mean   \n",
       "SK_ID_CURR                                                              \n",
       "100001                        3  0.428571                      0  0.0   \n",
       "100002                        2  0.250000                      0  0.0   \n",
       "100003                        1  0.250000                      0  0.0   \n",
       "100004                        0  0.000000                      0  0.0   \n",
       "100005                        2  0.666667                      0  0.0   \n",
       "\n",
       "           CREDIT_ACTIVE_Closed           CREDIT_ACTIVE_Sold       \\\n",
       "                            sum      mean                sum mean   \n",
       "SK_ID_CURR                                                          \n",
       "100001                        4  0.571429                  0  0.0   \n",
       "100002                        6  0.750000                  0  0.0   \n",
       "100003                        3  0.750000                  0  0.0   \n",
       "100004                        2  1.000000                  0  0.0   \n",
       "100005                        1  0.333333                  0  0.0   \n",
       "\n",
       "           CREDIT_CURRENCY_currency 1      ...  CREDIT_TYPE_Microloan       \\\n",
       "                                  sum mean ...                    sum mean   \n",
       "SK_ID_CURR                                 ...                               \n",
       "100001                              7  1.0 ...                      0  0.0   \n",
       "100002                              8  1.0 ...                      0  0.0   \n",
       "100003                              4  1.0 ...                      0  0.0   \n",
       "100004                              2  1.0 ...                      0  0.0   \n",
       "100005                              3  1.0 ...                      0  0.0   \n",
       "\n",
       "           CREDIT_TYPE_Mobile operator loan      CREDIT_TYPE_Mortgage       \\\n",
       "                                        sum mean                  sum mean   \n",
       "SK_ID_CURR                                                                   \n",
       "100001                                    0  0.0                    0  0.0   \n",
       "100002                                    0  0.0                    0  0.0   \n",
       "100003                                    0  0.0                    0  0.0   \n",
       "100004                                    0  0.0                    0  0.0   \n",
       "100005                                    0  0.0                    0  0.0   \n",
       "\n",
       "           CREDIT_TYPE_Real estate loan      CREDIT_TYPE_Unknown type of loan  \\\n",
       "                                    sum mean                              sum   \n",
       "SK_ID_CURR                                                                      \n",
       "100001                                0  0.0                                0   \n",
       "100002                                0  0.0                                0   \n",
       "100003                                0  0.0                                0   \n",
       "100004                                0  0.0                                0   \n",
       "100005                                0  0.0                                0   \n",
       "\n",
       "                 \n",
       "           mean  \n",
       "SK_ID_CURR       \n",
       "100001      0.0  \n",
       "100002      0.0  \n",
       "100003      0.0  \n",
       "100004      0.0  \n",
       "100005      0.0  \n",
       "\n",
       "[5 rows x 46 columns]"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "categorical_grouped = categorical.groupby('SK_ID_CURR').agg(['sum', 'mean'])\n",
    "categorical_grouped.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The `sum` columns represent the count of that category for the associated client and the `mean` represents the normalized count. One-hot encoding makes the process of calculating these figures very easy!\n",
    "\n",
    "We can use a similar function as before to rename the columns. Again, we have to deal with the multi-level index for the columns. We iterate through the first level (level 0) which is the name of the categorical variable appended with the value of the category (from one-hot encoding). Then we iterate  stats we calculated for each client. We will rename the column with the level 0 name appended with the stat. As an example, the column with `CREDIT_ACTIVE_Active` as level 0 and `sum` as level 1 will become `CREDIT_ACTIVE_Active_count`. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['CREDIT_ACTIVE_Active', 'CREDIT_ACTIVE_Bad debt',\n",
       "       'CREDIT_ACTIVE_Closed', 'CREDIT_ACTIVE_Sold',\n",
       "       'CREDIT_CURRENCY_currency 1', 'CREDIT_CURRENCY_currency 2',\n",
       "       'CREDIT_CURRENCY_currency 3', 'CREDIT_CURRENCY_currency 4',\n",
       "       'CREDIT_TYPE_Another type of loan', 'CREDIT_TYPE_Car loan'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "categorical_grouped.columns.levels[0][:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['sum', 'mean'], dtype='object')"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "categorical_grouped.columns.levels[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>CREDIT_ACTIVE_Active_count</th>\n",
       "      <th>CREDIT_ACTIVE_Active_count_norm</th>\n",
       "      <th>CREDIT_ACTIVE_Bad debt_count</th>\n",
       "      <th>CREDIT_ACTIVE_Bad debt_count_norm</th>\n",
       "      <th>CREDIT_ACTIVE_Closed_count</th>\n",
       "      <th>CREDIT_ACTIVE_Closed_count_norm</th>\n",
       "      <th>CREDIT_ACTIVE_Sold_count</th>\n",
       "      <th>CREDIT_ACTIVE_Sold_count_norm</th>\n",
       "      <th>CREDIT_CURRENCY_currency 1_count</th>\n",
       "      <th>CREDIT_CURRENCY_currency 1_count_norm</th>\n",
       "      <th>...</th>\n",
       "      <th>CREDIT_TYPE_Microloan_count</th>\n",
       "      <th>CREDIT_TYPE_Microloan_count_norm</th>\n",
       "      <th>CREDIT_TYPE_Mobile operator loan_count</th>\n",
       "      <th>CREDIT_TYPE_Mobile operator loan_count_norm</th>\n",
       "      <th>CREDIT_TYPE_Mortgage_count</th>\n",
       "      <th>CREDIT_TYPE_Mortgage_count_norm</th>\n",
       "      <th>CREDIT_TYPE_Real estate loan_count</th>\n",
       "      <th>CREDIT_TYPE_Real estate loan_count_norm</th>\n",
       "      <th>CREDIT_TYPE_Unknown type of loan_count</th>\n",
       "      <th>CREDIT_TYPE_Unknown type of loan_count_norm</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>100001</th>\n",
       "      <td>3</td>\n",
       "      <td>0.428571</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4</td>\n",
       "      <td>0.571429</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>7</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100002</th>\n",
       "      <td>2</td>\n",
       "      <td>0.250000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6</td>\n",
       "      <td>0.750000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>8</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100003</th>\n",
       "      <td>1</td>\n",
       "      <td>0.250000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3</td>\n",
       "      <td>0.750000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100004</th>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100005</th>\n",
       "      <td>2</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 46 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            CREDIT_ACTIVE_Active_count  CREDIT_ACTIVE_Active_count_norm  \\\n",
       "SK_ID_CURR                                                                \n",
       "100001                               3                         0.428571   \n",
       "100002                               2                         0.250000   \n",
       "100003                               1                         0.250000   \n",
       "100004                               0                         0.000000   \n",
       "100005                               2                         0.666667   \n",
       "\n",
       "            CREDIT_ACTIVE_Bad debt_count  CREDIT_ACTIVE_Bad debt_count_norm  \\\n",
       "SK_ID_CURR                                                                    \n",
       "100001                                 0                                0.0   \n",
       "100002                                 0                                0.0   \n",
       "100003                                 0                                0.0   \n",
       "100004                                 0                                0.0   \n",
       "100005                                 0                                0.0   \n",
       "\n",
       "            CREDIT_ACTIVE_Closed_count  CREDIT_ACTIVE_Closed_count_norm  \\\n",
       "SK_ID_CURR                                                                \n",
       "100001                               4                         0.571429   \n",
       "100002                               6                         0.750000   \n",
       "100003                               3                         0.750000   \n",
       "100004                               2                         1.000000   \n",
       "100005                               1                         0.333333   \n",
       "\n",
       "            CREDIT_ACTIVE_Sold_count  CREDIT_ACTIVE_Sold_count_norm  \\\n",
       "SK_ID_CURR                                                            \n",
       "100001                             0                            0.0   \n",
       "100002                             0                            0.0   \n",
       "100003                             0                            0.0   \n",
       "100004                             0                            0.0   \n",
       "100005                             0                            0.0   \n",
       "\n",
       "            CREDIT_CURRENCY_currency 1_count  \\\n",
       "SK_ID_CURR                                     \n",
       "100001                                     7   \n",
       "100002                                     8   \n",
       "100003                                     4   \n",
       "100004                                     2   \n",
       "100005                                     3   \n",
       "\n",
       "            CREDIT_CURRENCY_currency 1_count_norm  \\\n",
       "SK_ID_CURR                                          \n",
       "100001                                        1.0   \n",
       "100002                                        1.0   \n",
       "100003                                        1.0   \n",
       "100004                                        1.0   \n",
       "100005                                        1.0   \n",
       "\n",
       "                               ...                       \\\n",
       "SK_ID_CURR                     ...                        \n",
       "100001                         ...                        \n",
       "100002                         ...                        \n",
       "100003                         ...                        \n",
       "100004                         ...                        \n",
       "100005                         ...                        \n",
       "\n",
       "            CREDIT_TYPE_Microloan_count  CREDIT_TYPE_Microloan_count_norm  \\\n",
       "SK_ID_CURR                                                                  \n",
       "100001                                0                               0.0   \n",
       "100002                                0                               0.0   \n",
       "100003                                0                               0.0   \n",
       "100004                                0                               0.0   \n",
       "100005                                0                               0.0   \n",
       "\n",
       "            CREDIT_TYPE_Mobile operator loan_count  \\\n",
       "SK_ID_CURR                                           \n",
       "100001                                           0   \n",
       "100002                                           0   \n",
       "100003                                           0   \n",
       "100004                                           0   \n",
       "100005                                           0   \n",
       "\n",
       "            CREDIT_TYPE_Mobile operator loan_count_norm  \\\n",
       "SK_ID_CURR                                                \n",
       "100001                                              0.0   \n",
       "100002                                              0.0   \n",
       "100003                                              0.0   \n",
       "100004                                              0.0   \n",
       "100005                                              0.0   \n",
       "\n",
       "            CREDIT_TYPE_Mortgage_count  CREDIT_TYPE_Mortgage_count_norm  \\\n",
       "SK_ID_CURR                                                                \n",
       "100001                               0                              0.0   \n",
       "100002                               0                              0.0   \n",
       "100003                               0                              0.0   \n",
       "100004                               0                              0.0   \n",
       "100005                               0                              0.0   \n",
       "\n",
       "            CREDIT_TYPE_Real estate loan_count  \\\n",
       "SK_ID_CURR                                       \n",
       "100001                                       0   \n",
       "100002                                       0   \n",
       "100003                                       0   \n",
       "100004                                       0   \n",
       "100005                                       0   \n",
       "\n",
       "            CREDIT_TYPE_Real estate loan_count_norm  \\\n",
       "SK_ID_CURR                                            \n",
       "100001                                          0.0   \n",
       "100002                                          0.0   \n",
       "100003                                          0.0   \n",
       "100004                                          0.0   \n",
       "100005                                          0.0   \n",
       "\n",
       "            CREDIT_TYPE_Unknown type of loan_count  \\\n",
       "SK_ID_CURR                                           \n",
       "100001                                           0   \n",
       "100002                                           0   \n",
       "100003                                           0   \n",
       "100004                                           0   \n",
       "100005                                           0   \n",
       "\n",
       "            CREDIT_TYPE_Unknown type of loan_count_norm  \n",
       "SK_ID_CURR                                               \n",
       "100001                                              0.0  \n",
       "100002                                              0.0  \n",
       "100003                                              0.0  \n",
       "100004                                              0.0  \n",
       "100005                                              0.0  \n",
       "\n",
       "[5 rows x 46 columns]"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "group_var = 'SK_ID_CURR'\n",
    "\n",
    "# Need to create new column names\n",
    "columns = []\n",
    "\n",
    "# Iterate through the variables names\n",
    "for var in categorical_grouped.columns.levels[0]:\n",
    "    # Skip the grouping variable\n",
    "    if var != group_var:\n",
    "        # Iterate through the stat names\n",
    "        for stat in ['count', 'count_norm']:\n",
    "            # Make a new column name for the variable and stat\n",
    "            columns.append('%s_%s' % (var, stat))\n",
    "\n",
    "#  Rename the columns\n",
    "categorical_grouped.columns = columns\n",
    "\n",
    "categorical_grouped.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The sum column records the counts and the mean column records the normalized count. \n",
    "\n",
    "We can merge this dataframe into the training data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th>TARGET</th>\n",
       "      <th>NAME_CONTRACT_TYPE</th>\n",
       "      <th>CODE_GENDER</th>\n",
       "      <th>FLAG_OWN_CAR</th>\n",
       "      <th>FLAG_OWN_REALTY</th>\n",
       "      <th>CNT_CHILDREN</th>\n",
       "      <th>AMT_INCOME_TOTAL</th>\n",
       "      <th>AMT_CREDIT</th>\n",
       "      <th>AMT_ANNUITY</th>\n",
       "      <th>...</th>\n",
       "      <th>CREDIT_TYPE_Microloan_sum</th>\n",
       "      <th>CREDIT_TYPE_Microloan_mean</th>\n",
       "      <th>CREDIT_TYPE_Mobile operator loan_sum</th>\n",
       "      <th>CREDIT_TYPE_Mobile operator loan_mean</th>\n",
       "      <th>CREDIT_TYPE_Mortgage_sum</th>\n",
       "      <th>CREDIT_TYPE_Mortgage_mean</th>\n",
       "      <th>CREDIT_TYPE_Real estate loan_sum</th>\n",
       "      <th>CREDIT_TYPE_Real estate loan_mean</th>\n",
       "      <th>CREDIT_TYPE_Unknown type of loan_sum</th>\n",
       "      <th>CREDIT_TYPE_Unknown type of loan_mean</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>100002</td>\n",
       "      <td>1</td>\n",
       "      <td>Cash loans</td>\n",
       "      <td>M</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>202500.0</td>\n",
       "      <td>406597.5</td>\n",
       "      <td>24700.5</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>100003</td>\n",
       "      <td>0</td>\n",
       "      <td>Cash loans</td>\n",
       "      <td>F</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0</td>\n",
       "      <td>270000.0</td>\n",
       "      <td>1293502.5</td>\n",
       "      <td>35698.5</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>100004</td>\n",
       "      <td>0</td>\n",
       "      <td>Revolving loans</td>\n",
       "      <td>M</td>\n",
       "      <td>Y</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>67500.0</td>\n",
       "      <td>135000.0</td>\n",
       "      <td>6750.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>100006</td>\n",
       "      <td>0</td>\n",
       "      <td>Cash loans</td>\n",
       "      <td>F</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>135000.0</td>\n",
       "      <td>312682.5</td>\n",
       "      <td>29686.5</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>100007</td>\n",
       "      <td>0</td>\n",
       "      <td>Cash loans</td>\n",
       "      <td>M</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>121500.0</td>\n",
       "      <td>513000.0</td>\n",
       "      <td>21865.5</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 217 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_CURR  TARGET NAME_CONTRACT_TYPE CODE_GENDER FLAG_OWN_CAR  \\\n",
       "0      100002       1         Cash loans           M            N   \n",
       "1      100003       0         Cash loans           F            N   \n",
       "2      100004       0    Revolving loans           M            Y   \n",
       "3      100006       0         Cash loans           F            N   \n",
       "4      100007       0         Cash loans           M            N   \n",
       "\n",
       "  FLAG_OWN_REALTY  CNT_CHILDREN  AMT_INCOME_TOTAL  AMT_CREDIT  AMT_ANNUITY  \\\n",
       "0               Y             0          202500.0    406597.5      24700.5   \n",
       "1               N             0          270000.0   1293502.5      35698.5   \n",
       "2               Y             0           67500.0    135000.0       6750.0   \n",
       "3               Y             0          135000.0    312682.5      29686.5   \n",
       "4               Y             0          121500.0    513000.0      21865.5   \n",
       "\n",
       "                   ...                    CREDIT_TYPE_Microloan_sum  \\\n",
       "0                  ...                                          0.0   \n",
       "1                  ...                                          0.0   \n",
       "2                  ...                                          0.0   \n",
       "3                  ...                                          NaN   \n",
       "4                  ...                                          0.0   \n",
       "\n",
       "  CREDIT_TYPE_Microloan_mean CREDIT_TYPE_Mobile operator loan_sum  \\\n",
       "0                        0.0                                  0.0   \n",
       "1                        0.0                                  0.0   \n",
       "2                        0.0                                  0.0   \n",
       "3                        NaN                                  NaN   \n",
       "4                        0.0                                  0.0   \n",
       "\n",
       "  CREDIT_TYPE_Mobile operator loan_mean CREDIT_TYPE_Mortgage_sum  \\\n",
       "0                                   0.0                      0.0   \n",
       "1                                   0.0                      0.0   \n",
       "2                                   0.0                      0.0   \n",
       "3                                   NaN                      NaN   \n",
       "4                                   0.0                      0.0   \n",
       "\n",
       "  CREDIT_TYPE_Mortgage_mean  CREDIT_TYPE_Real estate loan_sum  \\\n",
       "0                       0.0                               0.0   \n",
       "1                       0.0                               0.0   \n",
       "2                       0.0                               0.0   \n",
       "3                       NaN                               NaN   \n",
       "4                       0.0                               0.0   \n",
       "\n",
       "   CREDIT_TYPE_Real estate loan_mean  CREDIT_TYPE_Unknown type of loan_sum  \\\n",
       "0                                0.0                                   0.0   \n",
       "1                                0.0                                   0.0   \n",
       "2                                0.0                                   0.0   \n",
       "3                                NaN                                   NaN   \n",
       "4                                0.0                                   0.0   \n",
       "\n",
       "   CREDIT_TYPE_Unknown type of loan_mean  \n",
       "0                                    0.0  \n",
       "1                                    0.0  \n",
       "2                                    0.0  \n",
       "3                                    NaN  \n",
       "4                                    0.0  \n",
       "\n",
       "[5 rows x 217 columns]"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train = train.merge(categorical_grouped, left_on = 'SK_ID_CURR', right_index = True, how = 'left')\n",
    "train.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "a30770ca960821a6c915f6acd54161a37bffb432"
   },
   "source": [
    "### Function to Handle Categorical Variables\n",
    "\n",
    "To make the code more efficient, we can now write a function to handle the categorical variables for us. This will take the same form as the `agg_numeric` function in that it accepts a dataframe and a grouping variable. Then it will calculate the counts and normalized counts of each category for all categorical variables in the dataframe."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {
    "_uuid": "ddc7decf9497eaa536255d601f76c3a6259fca6b",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def count_categorical(df, group_var, df_name):\n",
    "    \"\"\"Computes counts and normalized counts for each observation\n",
    "    of `group_var` of each unique category in every categorical variable\n",
    "    \n",
    "    Parameters\n",
    "    --------\n",
    "    df : dataframe \n",
    "        The dataframe to calculate the value counts for.\n",
    "        \n",
    "    group_var : string\n",
    "        The variable by which to group the dataframe. For each unique\n",
    "        value of this variable, the final dataframe will have one row\n",
    "        \n",
    "    df_name : string\n",
    "        Variable added to the front of column names to keep track of columns\n",
    "\n",
    "    \n",
    "    Return\n",
    "    --------\n",
    "    categorical : dataframe\n",
    "        A dataframe with counts and normalized counts of each unique category in every categorical variable\n",
    "        with one row for every unique value of the `group_var`.\n",
    "        \n",
    "    \"\"\"\n",
    "    \n",
    "    # Select the categorical columns\n",
    "    categorical = pd.get_dummies(df.select_dtypes('object'))\n",
    "\n",
    "    # Make sure to put the identifying id on the column\n",
    "    categorical[group_var] = df[group_var]\n",
    "\n",
    "    # Groupby the group var and calculate the sum and mean\n",
    "    categorical = categorical.groupby(group_var).agg(['sum', 'mean'])\n",
    "    \n",
    "    column_names = []\n",
    "    \n",
    "    # Iterate through the columns in level 0\n",
    "    for var in categorical.columns.levels[0]:\n",
    "        # Iterate through the stats in level 1\n",
    "        for stat in ['count', 'count_norm']:\n",
    "            # Make a new column name\n",
    "            column_names.append('%s_%s_%s' % (df_name, var, stat))\n",
    "    \n",
    "    categorical.columns = column_names\n",
    "    \n",
    "    return categorical"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {
    "_uuid": "d0f1629b07ec6e75e98da459f22611d275fcb282",
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>bureau_CREDIT_ACTIVE_Active_count</th>\n",
       "      <th>bureau_CREDIT_ACTIVE_Active_count_norm</th>\n",
       "      <th>bureau_CREDIT_ACTIVE_Bad debt_count</th>\n",
       "      <th>bureau_CREDIT_ACTIVE_Bad debt_count_norm</th>\n",
       "      <th>bureau_CREDIT_ACTIVE_Closed_count</th>\n",
       "      <th>bureau_CREDIT_ACTIVE_Closed_count_norm</th>\n",
       "      <th>bureau_CREDIT_ACTIVE_Sold_count</th>\n",
       "      <th>bureau_CREDIT_ACTIVE_Sold_count_norm</th>\n",
       "      <th>bureau_CREDIT_CURRENCY_currency 1_count</th>\n",
       "      <th>bureau_CREDIT_CURRENCY_currency 1_count_norm</th>\n",
       "      <th>...</th>\n",
       "      <th>bureau_CREDIT_TYPE_Microloan_count</th>\n",
       "      <th>bureau_CREDIT_TYPE_Microloan_count_norm</th>\n",
       "      <th>bureau_CREDIT_TYPE_Mobile operator loan_count</th>\n",
       "      <th>bureau_CREDIT_TYPE_Mobile operator loan_count_norm</th>\n",
       "      <th>bureau_CREDIT_TYPE_Mortgage_count</th>\n",
       "      <th>bureau_CREDIT_TYPE_Mortgage_count_norm</th>\n",
       "      <th>bureau_CREDIT_TYPE_Real estate loan_count</th>\n",
       "      <th>bureau_CREDIT_TYPE_Real estate loan_count_norm</th>\n",
       "      <th>bureau_CREDIT_TYPE_Unknown type of loan_count</th>\n",
       "      <th>bureau_CREDIT_TYPE_Unknown type of loan_count_norm</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>100001</th>\n",
       "      <td>3</td>\n",
       "      <td>0.428571</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4</td>\n",
       "      <td>0.571429</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>7</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100002</th>\n",
       "      <td>2</td>\n",
       "      <td>0.250000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6</td>\n",
       "      <td>0.750000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>8</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100003</th>\n",
       "      <td>1</td>\n",
       "      <td>0.250000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3</td>\n",
       "      <td>0.750000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100004</th>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100005</th>\n",
       "      <td>2</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 46 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            bureau_CREDIT_ACTIVE_Active_count  \\\n",
       "SK_ID_CURR                                      \n",
       "100001                                      3   \n",
       "100002                                      2   \n",
       "100003                                      1   \n",
       "100004                                      0   \n",
       "100005                                      2   \n",
       "\n",
       "            bureau_CREDIT_ACTIVE_Active_count_norm  \\\n",
       "SK_ID_CURR                                           \n",
       "100001                                    0.428571   \n",
       "100002                                    0.250000   \n",
       "100003                                    0.250000   \n",
       "100004                                    0.000000   \n",
       "100005                                    0.666667   \n",
       "\n",
       "            bureau_CREDIT_ACTIVE_Bad debt_count  \\\n",
       "SK_ID_CURR                                        \n",
       "100001                                        0   \n",
       "100002                                        0   \n",
       "100003                                        0   \n",
       "100004                                        0   \n",
       "100005                                        0   \n",
       "\n",
       "            bureau_CREDIT_ACTIVE_Bad debt_count_norm  \\\n",
       "SK_ID_CURR                                             \n",
       "100001                                           0.0   \n",
       "100002                                           0.0   \n",
       "100003                                           0.0   \n",
       "100004                                           0.0   \n",
       "100005                                           0.0   \n",
       "\n",
       "            bureau_CREDIT_ACTIVE_Closed_count  \\\n",
       "SK_ID_CURR                                      \n",
       "100001                                      4   \n",
       "100002                                      6   \n",
       "100003                                      3   \n",
       "100004                                      2   \n",
       "100005                                      1   \n",
       "\n",
       "            bureau_CREDIT_ACTIVE_Closed_count_norm  \\\n",
       "SK_ID_CURR                                           \n",
       "100001                                    0.571429   \n",
       "100002                                    0.750000   \n",
       "100003                                    0.750000   \n",
       "100004                                    1.000000   \n",
       "100005                                    0.333333   \n",
       "\n",
       "            bureau_CREDIT_ACTIVE_Sold_count  \\\n",
       "SK_ID_CURR                                    \n",
       "100001                                    0   \n",
       "100002                                    0   \n",
       "100003                                    0   \n",
       "100004                                    0   \n",
       "100005                                    0   \n",
       "\n",
       "            bureau_CREDIT_ACTIVE_Sold_count_norm  \\\n",
       "SK_ID_CURR                                         \n",
       "100001                                       0.0   \n",
       "100002                                       0.0   \n",
       "100003                                       0.0   \n",
       "100004                                       0.0   \n",
       "100005                                       0.0   \n",
       "\n",
       "            bureau_CREDIT_CURRENCY_currency 1_count  \\\n",
       "SK_ID_CURR                                            \n",
       "100001                                            7   \n",
       "100002                                            8   \n",
       "100003                                            4   \n",
       "100004                                            2   \n",
       "100005                                            3   \n",
       "\n",
       "            bureau_CREDIT_CURRENCY_currency 1_count_norm  \\\n",
       "SK_ID_CURR                                                 \n",
       "100001                                               1.0   \n",
       "100002                                               1.0   \n",
       "100003                                               1.0   \n",
       "100004                                               1.0   \n",
       "100005                                               1.0   \n",
       "\n",
       "                                   ...                          \\\n",
       "SK_ID_CURR                         ...                           \n",
       "100001                             ...                           \n",
       "100002                             ...                           \n",
       "100003                             ...                           \n",
       "100004                             ...                           \n",
       "100005                             ...                           \n",
       "\n",
       "            bureau_CREDIT_TYPE_Microloan_count  \\\n",
       "SK_ID_CURR                                       \n",
       "100001                                       0   \n",
       "100002                                       0   \n",
       "100003                                       0   \n",
       "100004                                       0   \n",
       "100005                                       0   \n",
       "\n",
       "            bureau_CREDIT_TYPE_Microloan_count_norm  \\\n",
       "SK_ID_CURR                                            \n",
       "100001                                          0.0   \n",
       "100002                                          0.0   \n",
       "100003                                          0.0   \n",
       "100004                                          0.0   \n",
       "100005                                          0.0   \n",
       "\n",
       "            bureau_CREDIT_TYPE_Mobile operator loan_count  \\\n",
       "SK_ID_CURR                                                  \n",
       "100001                                                  0   \n",
       "100002                                                  0   \n",
       "100003                                                  0   \n",
       "100004                                                  0   \n",
       "100005                                                  0   \n",
       "\n",
       "            bureau_CREDIT_TYPE_Mobile operator loan_count_norm  \\\n",
       "SK_ID_CURR                                                       \n",
       "100001                                                    0.0    \n",
       "100002                                                    0.0    \n",
       "100003                                                    0.0    \n",
       "100004                                                    0.0    \n",
       "100005                                                    0.0    \n",
       "\n",
       "            bureau_CREDIT_TYPE_Mortgage_count  \\\n",
       "SK_ID_CURR                                      \n",
       "100001                                      0   \n",
       "100002                                      0   \n",
       "100003                                      0   \n",
       "100004                                      0   \n",
       "100005                                      0   \n",
       "\n",
       "            bureau_CREDIT_TYPE_Mortgage_count_norm  \\\n",
       "SK_ID_CURR                                           \n",
       "100001                                         0.0   \n",
       "100002                                         0.0   \n",
       "100003                                         0.0   \n",
       "100004                                         0.0   \n",
       "100005                                         0.0   \n",
       "\n",
       "            bureau_CREDIT_TYPE_Real estate loan_count  \\\n",
       "SK_ID_CURR                                              \n",
       "100001                                              0   \n",
       "100002                                              0   \n",
       "100003                                              0   \n",
       "100004                                              0   \n",
       "100005                                              0   \n",
       "\n",
       "            bureau_CREDIT_TYPE_Real estate loan_count_norm  \\\n",
       "SK_ID_CURR                                                   \n",
       "100001                                                 0.0   \n",
       "100002                                                 0.0   \n",
       "100003                                                 0.0   \n",
       "100004                                                 0.0   \n",
       "100005                                                 0.0   \n",
       "\n",
       "            bureau_CREDIT_TYPE_Unknown type of loan_count  \\\n",
       "SK_ID_CURR                                                  \n",
       "100001                                                  0   \n",
       "100002                                                  0   \n",
       "100003                                                  0   \n",
       "100004                                                  0   \n",
       "100005                                                  0   \n",
       "\n",
       "            bureau_CREDIT_TYPE_Unknown type of loan_count_norm  \n",
       "SK_ID_CURR                                                      \n",
       "100001                                                    0.0   \n",
       "100002                                                    0.0   \n",
       "100003                                                    0.0   \n",
       "100004                                                    0.0   \n",
       "100005                                                    0.0   \n",
       "\n",
       "[5 rows x 46 columns]"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bureau_counts = count_categorical(bureau, group_var = 'SK_ID_CURR', df_name = 'bureau')\n",
    "bureau_counts.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "cecd442561b3a8405721875878bcbb90c94f2517"
   },
   "source": [
    "### Applying Operations to another dataframe\n",
    "\n",
    "We will now turn to the bureau balance dataframe. This dataframe has monthly information about each client's previous loan(s) with other financial institutions. Instead of grouping this dataframe by the `SK_ID_CURR` which is the client id, we will first group the dataframe by the `SK_ID_BUREAU` which is the id of the previous loan. This will give us one row of the dataframe for each loan. Then, we can group by the `SK_ID_CURR` and calculate the aggregations across the loans of each client. The final result will be a dataframe with one row for each client, with stats calculated for their loans."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {
    "_uuid": "b49d1c1553eb3c0a8887d967fe9c981b015a7e40",
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_BUREAU</th>\n",
       "      <th>MONTHS_BALANCE</th>\n",
       "      <th>STATUS</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5715448</td>\n",
       "      <td>0</td>\n",
       "      <td>C</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5715448</td>\n",
       "      <td>-1</td>\n",
       "      <td>C</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>5715448</td>\n",
       "      <td>-2</td>\n",
       "      <td>C</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>5715448</td>\n",
       "      <td>-3</td>\n",
       "      <td>C</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5715448</td>\n",
       "      <td>-4</td>\n",
       "      <td>C</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_BUREAU  MONTHS_BALANCE STATUS\n",
       "0       5715448               0      C\n",
       "1       5715448              -1      C\n",
       "2       5715448              -2      C\n",
       "3       5715448              -3      C\n",
       "4       5715448              -4      C"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Read in bureau balance\n",
    "bureau_balance = pd.read_csv('../input/bureau_balance.csv')\n",
    "bureau_balance.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "815ea915661653e3322e5cbf0f60614754411526"
   },
   "source": [
    "First, we can calculate the value counts of each status for each loan. Fortunately, we already have a function that does this for us! "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {
    "_uuid": "891d69c6cf3634d2fab71b0c8cc1eb212f433083",
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>bureau_balance_STATUS_0_count</th>\n",
       "      <th>bureau_balance_STATUS_0_count_norm</th>\n",
       "      <th>bureau_balance_STATUS_1_count</th>\n",
       "      <th>bureau_balance_STATUS_1_count_norm</th>\n",
       "      <th>bureau_balance_STATUS_2_count</th>\n",
       "      <th>bureau_balance_STATUS_2_count_norm</th>\n",
       "      <th>bureau_balance_STATUS_3_count</th>\n",
       "      <th>bureau_balance_STATUS_3_count_norm</th>\n",
       "      <th>bureau_balance_STATUS_4_count</th>\n",
       "      <th>bureau_balance_STATUS_4_count_norm</th>\n",
       "      <th>bureau_balance_STATUS_5_count</th>\n",
       "      <th>bureau_balance_STATUS_5_count_norm</th>\n",
       "      <th>bureau_balance_STATUS_C_count</th>\n",
       "      <th>bureau_balance_STATUS_C_count_norm</th>\n",
       "      <th>bureau_balance_STATUS_X_count</th>\n",
       "      <th>bureau_balance_STATUS_X_count_norm</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_ID_BUREAU</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>5001709</th>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>86</td>\n",
       "      <td>0.886598</td>\n",
       "      <td>11</td>\n",
       "      <td>0.113402</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5001710</th>\n",
       "      <td>5</td>\n",
       "      <td>0.060241</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>48</td>\n",
       "      <td>0.578313</td>\n",
       "      <td>30</td>\n",
       "      <td>0.361446</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5001711</th>\n",
       "      <td>3</td>\n",
       "      <td>0.750000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1</td>\n",
       "      <td>0.250000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5001712</th>\n",
       "      <td>10</td>\n",
       "      <td>0.526316</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>9</td>\n",
       "      <td>0.473684</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5001713</th>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>22</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              bureau_balance_STATUS_0_count  \\\n",
       "SK_ID_BUREAU                                  \n",
       "5001709                                   0   \n",
       "5001710                                   5   \n",
       "5001711                                   3   \n",
       "5001712                                  10   \n",
       "5001713                                   0   \n",
       "\n",
       "              bureau_balance_STATUS_0_count_norm  \\\n",
       "SK_ID_BUREAU                                       \n",
       "5001709                                 0.000000   \n",
       "5001710                                 0.060241   \n",
       "5001711                                 0.750000   \n",
       "5001712                                 0.526316   \n",
       "5001713                                 0.000000   \n",
       "\n",
       "              bureau_balance_STATUS_1_count  \\\n",
       "SK_ID_BUREAU                                  \n",
       "5001709                                   0   \n",
       "5001710                                   0   \n",
       "5001711                                   0   \n",
       "5001712                                   0   \n",
       "5001713                                   0   \n",
       "\n",
       "              bureau_balance_STATUS_1_count_norm  \\\n",
       "SK_ID_BUREAU                                       \n",
       "5001709                                      0.0   \n",
       "5001710                                      0.0   \n",
       "5001711                                      0.0   \n",
       "5001712                                      0.0   \n",
       "5001713                                      0.0   \n",
       "\n",
       "              bureau_balance_STATUS_2_count  \\\n",
       "SK_ID_BUREAU                                  \n",
       "5001709                                   0   \n",
       "5001710                                   0   \n",
       "5001711                                   0   \n",
       "5001712                                   0   \n",
       "5001713                                   0   \n",
       "\n",
       "              bureau_balance_STATUS_2_count_norm  \\\n",
       "SK_ID_BUREAU                                       \n",
       "5001709                                      0.0   \n",
       "5001710                                      0.0   \n",
       "5001711                                      0.0   \n",
       "5001712                                      0.0   \n",
       "5001713                                      0.0   \n",
       "\n",
       "              bureau_balance_STATUS_3_count  \\\n",
       "SK_ID_BUREAU                                  \n",
       "5001709                                   0   \n",
       "5001710                                   0   \n",
       "5001711                                   0   \n",
       "5001712                                   0   \n",
       "5001713                                   0   \n",
       "\n",
       "              bureau_balance_STATUS_3_count_norm  \\\n",
       "SK_ID_BUREAU                                       \n",
       "5001709                                      0.0   \n",
       "5001710                                      0.0   \n",
       "5001711                                      0.0   \n",
       "5001712                                      0.0   \n",
       "5001713                                      0.0   \n",
       "\n",
       "              bureau_balance_STATUS_4_count  \\\n",
       "SK_ID_BUREAU                                  \n",
       "5001709                                   0   \n",
       "5001710                                   0   \n",
       "5001711                                   0   \n",
       "5001712                                   0   \n",
       "5001713                                   0   \n",
       "\n",
       "              bureau_balance_STATUS_4_count_norm  \\\n",
       "SK_ID_BUREAU                                       \n",
       "5001709                                      0.0   \n",
       "5001710                                      0.0   \n",
       "5001711                                      0.0   \n",
       "5001712                                      0.0   \n",
       "5001713                                      0.0   \n",
       "\n",
       "              bureau_balance_STATUS_5_count  \\\n",
       "SK_ID_BUREAU                                  \n",
       "5001709                                   0   \n",
       "5001710                                   0   \n",
       "5001711                                   0   \n",
       "5001712                                   0   \n",
       "5001713                                   0   \n",
       "\n",
       "              bureau_balance_STATUS_5_count_norm  \\\n",
       "SK_ID_BUREAU                                       \n",
       "5001709                                      0.0   \n",
       "5001710                                      0.0   \n",
       "5001711                                      0.0   \n",
       "5001712                                      0.0   \n",
       "5001713                                      0.0   \n",
       "\n",
       "              bureau_balance_STATUS_C_count  \\\n",
       "SK_ID_BUREAU                                  \n",
       "5001709                                  86   \n",
       "5001710                                  48   \n",
       "5001711                                   0   \n",
       "5001712                                   9   \n",
       "5001713                                   0   \n",
       "\n",
       "              bureau_balance_STATUS_C_count_norm  \\\n",
       "SK_ID_BUREAU                                       \n",
       "5001709                                 0.886598   \n",
       "5001710                                 0.578313   \n",
       "5001711                                 0.000000   \n",
       "5001712                                 0.473684   \n",
       "5001713                                 0.000000   \n",
       "\n",
       "              bureau_balance_STATUS_X_count  \\\n",
       "SK_ID_BUREAU                                  \n",
       "5001709                                  11   \n",
       "5001710                                  30   \n",
       "5001711                                   1   \n",
       "5001712                                   0   \n",
       "5001713                                  22   \n",
       "\n",
       "              bureau_balance_STATUS_X_count_norm  \n",
       "SK_ID_BUREAU                                      \n",
       "5001709                                 0.113402  \n",
       "5001710                                 0.361446  \n",
       "5001711                                 0.250000  \n",
       "5001712                                 0.000000  \n",
       "5001713                                 1.000000  "
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Counts of each type of status for each previous loan\n",
    "bureau_balance_counts = count_categorical(bureau_balance, group_var = 'SK_ID_BUREAU', df_name = 'bureau_balance')\n",
    "bureau_balance_counts.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "4b1b1623a426efd45f7488636ffffc6b54c68d8e"
   },
   "source": [
    "Now we can handle the one numeric column. The `MONTHS_BALANCE` column has the \"months of balance relative to application date.\" This might not necessarily be that important as a numeric variable, and in future work we might want to consider this as a time variable. For now, we can just calculate the same aggregation statistics as previously. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {
    "_uuid": "da531864649dc2b8dab18c31019fc36d6ec6469d",
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_BUREAU</th>\n",
       "      <th>bureau_balance_MONTHS_BALANCE_mean</th>\n",
       "      <th>bureau_balance_MONTHS_BALANCE_max</th>\n",
       "      <th>bureau_balance_MONTHS_BALANCE_min</th>\n",
       "      <th>bureau_balance_MONTHS_BALANCE_sum</th>\n",
       "      <th>bureau_balance_counts</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5001709</td>\n",
       "      <td>-48.0</td>\n",
       "      <td>0</td>\n",
       "      <td>-96</td>\n",
       "      <td>-4656</td>\n",
       "      <td>97</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5001710</td>\n",
       "      <td>-41.0</td>\n",
       "      <td>0</td>\n",
       "      <td>-82</td>\n",
       "      <td>-3403</td>\n",
       "      <td>83</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>5001711</td>\n",
       "      <td>-1.5</td>\n",
       "      <td>0</td>\n",
       "      <td>-3</td>\n",
       "      <td>-6</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>5001712</td>\n",
       "      <td>-9.0</td>\n",
       "      <td>0</td>\n",
       "      <td>-18</td>\n",
       "      <td>-171</td>\n",
       "      <td>19</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5001713</td>\n",
       "      <td>-10.5</td>\n",
       "      <td>0</td>\n",
       "      <td>-21</td>\n",
       "      <td>-231</td>\n",
       "      <td>22</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_BUREAU  bureau_balance_MONTHS_BALANCE_mean  \\\n",
       "0       5001709                               -48.0   \n",
       "1       5001710                               -41.0   \n",
       "2       5001711                                -1.5   \n",
       "3       5001712                                -9.0   \n",
       "4       5001713                               -10.5   \n",
       "\n",
       "   bureau_balance_MONTHS_BALANCE_max  bureau_balance_MONTHS_BALANCE_min  \\\n",
       "0                                  0                                -96   \n",
       "1                                  0                                -82   \n",
       "2                                  0                                 -3   \n",
       "3                                  0                                -18   \n",
       "4                                  0                                -21   \n",
       "\n",
       "   bureau_balance_MONTHS_BALANCE_sum  bureau_balance_counts  \n",
       "0                              -4656                     97  \n",
       "1                              -3403                     83  \n",
       "2                                 -6                      4  \n",
       "3                               -171                     19  \n",
       "4                               -231                     22  "
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Calculate value count statistics for each `SK_ID_CURR` \n",
    "bureau_balance_agg = agg_numeric(bureau_balance, group_var = 'SK_ID_BUREAU', df_name = 'bureau_balance')\n",
    "bureau_balance_agg.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "50b714b2165463138c29715ad68919bcc47aac2d"
   },
   "source": [
    "The above dataframes have the calculations done on each _loan_. Now we need to aggregate these for each _client_. We can do this by merging the dataframes together first and then since all the variables are numeric, we just need to aggregate the statistics again, this time grouping by the `SK_ID_CURR`. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {
    "_uuid": "dbd657d3cbe3cf9948ba8d7543f6653e61025bc6",
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_BUREAU</th>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th>bureau_balance_MONTHS_BALANCE_mean</th>\n",
       "      <th>bureau_balance_MONTHS_BALANCE_max</th>\n",
       "      <th>bureau_balance_MONTHS_BALANCE_min</th>\n",
       "      <th>bureau_balance_MONTHS_BALANCE_sum</th>\n",
       "      <th>bureau_balance_counts</th>\n",
       "      <th>bureau_balance_STATUS_0_count</th>\n",
       "      <th>bureau_balance_STATUS_0_count_norm</th>\n",
       "      <th>bureau_balance_STATUS_1_count</th>\n",
       "      <th>...</th>\n",
       "      <th>bureau_balance_STATUS_3_count</th>\n",
       "      <th>bureau_balance_STATUS_3_count_norm</th>\n",
       "      <th>bureau_balance_STATUS_4_count</th>\n",
       "      <th>bureau_balance_STATUS_4_count_norm</th>\n",
       "      <th>bureau_balance_STATUS_5_count</th>\n",
       "      <th>bureau_balance_STATUS_5_count_norm</th>\n",
       "      <th>bureau_balance_STATUS_C_count</th>\n",
       "      <th>bureau_balance_STATUS_C_count_norm</th>\n",
       "      <th>bureau_balance_STATUS_X_count</th>\n",
       "      <th>bureau_balance_STATUS_X_count_norm</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5714462</td>\n",
       "      <td>215354</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5714463</td>\n",
       "      <td>215354</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>5714464</td>\n",
       "      <td>215354</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>5714465</td>\n",
       "      <td>215354</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5714466</td>\n",
       "      <td>215354</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 23 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_BUREAU  SK_ID_CURR  bureau_balance_MONTHS_BALANCE_mean  \\\n",
       "0       5714462      215354                                 NaN   \n",
       "1       5714463      215354                                 NaN   \n",
       "2       5714464      215354                                 NaN   \n",
       "3       5714465      215354                                 NaN   \n",
       "4       5714466      215354                                 NaN   \n",
       "\n",
       "   bureau_balance_MONTHS_BALANCE_max  bureau_balance_MONTHS_BALANCE_min  \\\n",
       "0                                NaN                                NaN   \n",
       "1                                NaN                                NaN   \n",
       "2                                NaN                                NaN   \n",
       "3                                NaN                                NaN   \n",
       "4                                NaN                                NaN   \n",
       "\n",
       "   bureau_balance_MONTHS_BALANCE_sum  bureau_balance_counts  \\\n",
       "0                                NaN                    NaN   \n",
       "1                                NaN                    NaN   \n",
       "2                                NaN                    NaN   \n",
       "3                                NaN                    NaN   \n",
       "4                                NaN                    NaN   \n",
       "\n",
       "   bureau_balance_STATUS_0_count  bureau_balance_STATUS_0_count_norm  \\\n",
       "0                            NaN                                 NaN   \n",
       "1                            NaN                                 NaN   \n",
       "2                            NaN                                 NaN   \n",
       "3                            NaN                                 NaN   \n",
       "4                            NaN                                 NaN   \n",
       "\n",
       "   bureau_balance_STATUS_1_count                 ...                  \\\n",
       "0                            NaN                 ...                   \n",
       "1                            NaN                 ...                   \n",
       "2                            NaN                 ...                   \n",
       "3                            NaN                 ...                   \n",
       "4                            NaN                 ...                   \n",
       "\n",
       "   bureau_balance_STATUS_3_count  bureau_balance_STATUS_3_count_norm  \\\n",
       "0                            NaN                                 NaN   \n",
       "1                            NaN                                 NaN   \n",
       "2                            NaN                                 NaN   \n",
       "3                            NaN                                 NaN   \n",
       "4                            NaN                                 NaN   \n",
       "\n",
       "   bureau_balance_STATUS_4_count  bureau_balance_STATUS_4_count_norm  \\\n",
       "0                            NaN                                 NaN   \n",
       "1                            NaN                                 NaN   \n",
       "2                            NaN                                 NaN   \n",
       "3                            NaN                                 NaN   \n",
       "4                            NaN                                 NaN   \n",
       "\n",
       "   bureau_balance_STATUS_5_count  bureau_balance_STATUS_5_count_norm  \\\n",
       "0                            NaN                                 NaN   \n",
       "1                            NaN                                 NaN   \n",
       "2                            NaN                                 NaN   \n",
       "3                            NaN                                 NaN   \n",
       "4                            NaN                                 NaN   \n",
       "\n",
       "   bureau_balance_STATUS_C_count  bureau_balance_STATUS_C_count_norm  \\\n",
       "0                            NaN                                 NaN   \n",
       "1                            NaN                                 NaN   \n",
       "2                            NaN                                 NaN   \n",
       "3                            NaN                                 NaN   \n",
       "4                            NaN                                 NaN   \n",
       "\n",
       "   bureau_balance_STATUS_X_count  bureau_balance_STATUS_X_count_norm  \n",
       "0                            NaN                                 NaN  \n",
       "1                            NaN                                 NaN  \n",
       "2                            NaN                                 NaN  \n",
       "3                            NaN                                 NaN  \n",
       "4                            NaN                                 NaN  \n",
       "\n",
       "[5 rows x 23 columns]"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Dataframe grouped by the loan\n",
    "bureau_by_loan = bureau_balance_agg.merge(bureau_balance_counts, right_index = True, left_on = 'SK_ID_BUREAU', how = 'outer')\n",
    "\n",
    "# Merge to include the SK_ID_CURR\n",
    "bureau_by_loan = bureau[['SK_ID_BUREAU', 'SK_ID_CURR']].merge(bureau_by_loan, on = 'SK_ID_BUREAU', how = 'left')\n",
    "\n",
    "bureau_by_loan.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {
    "_uuid": "ae22ded2231b37858d9d1d0c3c5b0eb72202c6ab",
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th>client_bureau_balance_MONTHS_BALANCE_mean_mean</th>\n",
       "      <th>client_bureau_balance_MONTHS_BALANCE_mean_max</th>\n",
       "      <th>client_bureau_balance_MONTHS_BALANCE_mean_min</th>\n",
       "      <th>client_bureau_balance_MONTHS_BALANCE_mean_sum</th>\n",
       "      <th>client_bureau_balance_MONTHS_BALANCE_max_mean</th>\n",
       "      <th>client_bureau_balance_MONTHS_BALANCE_max_max</th>\n",
       "      <th>client_bureau_balance_MONTHS_BALANCE_max_min</th>\n",
       "      <th>client_bureau_balance_MONTHS_BALANCE_max_sum</th>\n",
       "      <th>client_bureau_balance_MONTHS_BALANCE_min_mean</th>\n",
       "      <th>...</th>\n",
       "      <th>client_bureau_balance_STATUS_C_count_norm_sum</th>\n",
       "      <th>client_bureau_balance_STATUS_X_count_mean</th>\n",
       "      <th>client_bureau_balance_STATUS_X_count_max</th>\n",
       "      <th>client_bureau_balance_STATUS_X_count_min</th>\n",
       "      <th>client_bureau_balance_STATUS_X_count_sum</th>\n",
       "      <th>client_bureau_balance_STATUS_X_count_norm_mean</th>\n",
       "      <th>client_bureau_balance_STATUS_X_count_norm_max</th>\n",
       "      <th>client_bureau_balance_STATUS_X_count_norm_min</th>\n",
       "      <th>client_bureau_balance_STATUS_X_count_norm_sum</th>\n",
       "      <th>client_counts</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>100001</td>\n",
       "      <td>-11.785714</td>\n",
       "      <td>-0.5</td>\n",
       "      <td>-25.5</td>\n",
       "      <td>-82.5</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-23.571429</td>\n",
       "      <td>...</td>\n",
       "      <td>3.088683</td>\n",
       "      <td>4.285714</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>30.0</td>\n",
       "      <td>0.214590</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.502129</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>100002</td>\n",
       "      <td>-21.875000</td>\n",
       "      <td>-1.5</td>\n",
       "      <td>-39.5</td>\n",
       "      <td>-175.0</td>\n",
       "      <td>-15.5</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-32.0</td>\n",
       "      <td>-124.0</td>\n",
       "      <td>-28.250000</td>\n",
       "      <td>...</td>\n",
       "      <td>1.403409</td>\n",
       "      <td>1.875000</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>0.161932</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.295455</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>100003</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>100004</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>100005</td>\n",
       "      <td>-3.000000</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>-6.0</td>\n",
       "      <td>-9.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-6.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.384615</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.136752</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.410256</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 86 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_CURR  client_bureau_balance_MONTHS_BALANCE_mean_mean  \\\n",
       "0      100001                                      -11.785714   \n",
       "1      100002                                      -21.875000   \n",
       "2      100003                                             NaN   \n",
       "3      100004                                             NaN   \n",
       "4      100005                                       -3.000000   \n",
       "\n",
       "   client_bureau_balance_MONTHS_BALANCE_mean_max  \\\n",
       "0                                           -0.5   \n",
       "1                                           -1.5   \n",
       "2                                            NaN   \n",
       "3                                            NaN   \n",
       "4                                           -1.0   \n",
       "\n",
       "   client_bureau_balance_MONTHS_BALANCE_mean_min  \\\n",
       "0                                          -25.5   \n",
       "1                                          -39.5   \n",
       "2                                            NaN   \n",
       "3                                            NaN   \n",
       "4                                           -6.0   \n",
       "\n",
       "   client_bureau_balance_MONTHS_BALANCE_mean_sum  \\\n",
       "0                                          -82.5   \n",
       "1                                         -175.0   \n",
       "2                                            0.0   \n",
       "3                                            0.0   \n",
       "4                                           -9.0   \n",
       "\n",
       "   client_bureau_balance_MONTHS_BALANCE_max_mean  \\\n",
       "0                                            0.0   \n",
       "1                                          -15.5   \n",
       "2                                            NaN   \n",
       "3                                            NaN   \n",
       "4                                            0.0   \n",
       "\n",
       "   client_bureau_balance_MONTHS_BALANCE_max_max  \\\n",
       "0                                           0.0   \n",
       "1                                           0.0   \n",
       "2                                           NaN   \n",
       "3                                           NaN   \n",
       "4                                           0.0   \n",
       "\n",
       "   client_bureau_balance_MONTHS_BALANCE_max_min  \\\n",
       "0                                           0.0   \n",
       "1                                         -32.0   \n",
       "2                                           NaN   \n",
       "3                                           NaN   \n",
       "4                                           0.0   \n",
       "\n",
       "   client_bureau_balance_MONTHS_BALANCE_max_sum  \\\n",
       "0                                           0.0   \n",
       "1                                        -124.0   \n",
       "2                                           0.0   \n",
       "3                                           0.0   \n",
       "4                                           0.0   \n",
       "\n",
       "   client_bureau_balance_MONTHS_BALANCE_min_mean      ...        \\\n",
       "0                                     -23.571429      ...         \n",
       "1                                     -28.250000      ...         \n",
       "2                                            NaN      ...         \n",
       "3                                            NaN      ...         \n",
       "4                                      -6.000000      ...         \n",
       "\n",
       "   client_bureau_balance_STATUS_C_count_norm_sum  \\\n",
       "0                                       3.088683   \n",
       "1                                       1.403409   \n",
       "2                                       0.000000   \n",
       "3                                       0.000000   \n",
       "4                                       0.384615   \n",
       "\n",
       "   client_bureau_balance_STATUS_X_count_mean  \\\n",
       "0                                   4.285714   \n",
       "1                                   1.875000   \n",
       "2                                        NaN   \n",
       "3                                        NaN   \n",
       "4                                   0.666667   \n",
       "\n",
       "   client_bureau_balance_STATUS_X_count_max  \\\n",
       "0                                       9.0   \n",
       "1                                       3.0   \n",
       "2                                       NaN   \n",
       "3                                       NaN   \n",
       "4                                       1.0   \n",
       "\n",
       "   client_bureau_balance_STATUS_X_count_min  \\\n",
       "0                                       0.0   \n",
       "1                                       0.0   \n",
       "2                                       NaN   \n",
       "3                                       NaN   \n",
       "4                                       0.0   \n",
       "\n",
       "   client_bureau_balance_STATUS_X_count_sum  \\\n",
       "0                                      30.0   \n",
       "1                                      15.0   \n",
       "2                                       0.0   \n",
       "3                                       0.0   \n",
       "4                                       2.0   \n",
       "\n",
       "   client_bureau_balance_STATUS_X_count_norm_mean  \\\n",
       "0                                        0.214590   \n",
       "1                                        0.161932   \n",
       "2                                             NaN   \n",
       "3                                             NaN   \n",
       "4                                        0.136752   \n",
       "\n",
       "   client_bureau_balance_STATUS_X_count_norm_max  \\\n",
       "0                                       0.500000   \n",
       "1                                       0.500000   \n",
       "2                                            NaN   \n",
       "3                                            NaN   \n",
       "4                                       0.333333   \n",
       "\n",
       "   client_bureau_balance_STATUS_X_count_norm_min  \\\n",
       "0                                            0.0   \n",
       "1                                            0.0   \n",
       "2                                            NaN   \n",
       "3                                            NaN   \n",
       "4                                            0.0   \n",
       "\n",
       "   client_bureau_balance_STATUS_X_count_norm_sum  client_counts  \n",
       "0                                       1.502129              7  \n",
       "1                                       1.295455              8  \n",
       "2                                       0.000000              0  \n",
       "3                                       0.000000              0  \n",
       "4                                       0.410256              3  \n",
       "\n",
       "[5 rows x 86 columns]"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bureau_balance_by_client = agg_numeric(bureau_by_loan.drop(columns = ['SK_ID_BUREAU']), group_var = 'SK_ID_CURR', df_name = 'client')\n",
    "bureau_balance_by_client.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "fb037a66b921c2dbf837e8d6ddb5179be51e8446"
   },
   "source": [
    "To recap, for the `bureau_balance` dataframe we:\n",
    "\n",
    "1. Calculated numeric stats grouping by each loan\n",
    "2. Made value counts of each categorical variable grouping by loan\n",
    "3. Merged the stats and the value counts on the loans\n",
    "4. Calculated numeric stats for the resulting dataframe grouping by the client id\n",
    "\n",
    "The final resulting dataframe has one row for each client, with statistics calculated for all of their loans with monthly balance information. \n",
    "\n",
    "Some of these variables are a little confusing, so let's try to explain a few:\n",
    "\n",
    "* `client_bureau_balance_MONTHS_BALANCE_mean_mean`: For each loan calculate the mean value of `MONTHS_BALANCE`. Then for each client, calculate the mean of this value for all of their loans. \n",
    "* `client_bureau_balance_STATUS_X_count_norm_sum`: For each loan, calculate the number of occurences of `STATUS` == X divided by the number of total `STATUS` values for the loan. Then, for each client, add up the values for each loan. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "e7c7be26bc22cacc8168f907d01c2c259ad6c867"
   },
   "source": [
    "We will hold off on calculating the correlations until we have all the variables together in one dataframe. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "90aa8e57fbff00777a1e62b060c2c06cf7845381"
   },
   "source": [
    "# Putting the Functions Together\n",
    "\n",
    "We now have all the pieces in place to take the information from the previous loans at other institutions and the monthly payments information about these loans and put them into the main training dataframe. Let's do a reset of all the variables and then use the functions we built to do this from the ground up. This demonstrate the benefit of using functions for repeatable workflows! "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {
    "_uuid": "26848bd0252a31559e254c454b905305701ee522",
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "6226"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Free up memory by deleting old objects\n",
    "import gc\n",
    "gc.enable()\n",
    "del train, bureau, bureau_balance, bureau_agg, bureau_agg_new, bureau_balance_agg, bureau_balance_counts, bureau_by_loan, bureau_balance_by_client, bureau_counts\n",
    "gc.collect()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {
    "_uuid": "dc574b9486fb40f72146e6481b341931dd180e3d",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Read in new copies of all the dataframes\n",
    "train = pd.read_csv('../input/application_train.csv')\n",
    "bureau = pd.read_csv('../input/bureau.csv')\n",
    "bureau_balance = pd.read_csv('../input/bureau_balance.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "feba956fa449260cc71505cb9156887a62ca8337"
   },
   "source": [
    "### Counts of Bureau Dataframe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {
    "_uuid": "491f45986682beb4ee2bd62c32c622830b3b0e76",
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>bureau_CREDIT_ACTIVE_Active_count</th>\n",
       "      <th>bureau_CREDIT_ACTIVE_Active_count_norm</th>\n",
       "      <th>bureau_CREDIT_ACTIVE_Bad debt_count</th>\n",
       "      <th>bureau_CREDIT_ACTIVE_Bad debt_count_norm</th>\n",
       "      <th>bureau_CREDIT_ACTIVE_Closed_count</th>\n",
       "      <th>bureau_CREDIT_ACTIVE_Closed_count_norm</th>\n",
       "      <th>bureau_CREDIT_ACTIVE_Sold_count</th>\n",
       "      <th>bureau_CREDIT_ACTIVE_Sold_count_norm</th>\n",
       "      <th>bureau_CREDIT_CURRENCY_currency 1_count</th>\n",
       "      <th>bureau_CREDIT_CURRENCY_currency 1_count_norm</th>\n",
       "      <th>...</th>\n",
       "      <th>bureau_CREDIT_TYPE_Microloan_count</th>\n",
       "      <th>bureau_CREDIT_TYPE_Microloan_count_norm</th>\n",
       "      <th>bureau_CREDIT_TYPE_Mobile operator loan_count</th>\n",
       "      <th>bureau_CREDIT_TYPE_Mobile operator loan_count_norm</th>\n",
       "      <th>bureau_CREDIT_TYPE_Mortgage_count</th>\n",
       "      <th>bureau_CREDIT_TYPE_Mortgage_count_norm</th>\n",
       "      <th>bureau_CREDIT_TYPE_Real estate loan_count</th>\n",
       "      <th>bureau_CREDIT_TYPE_Real estate loan_count_norm</th>\n",
       "      <th>bureau_CREDIT_TYPE_Unknown type of loan_count</th>\n",
       "      <th>bureau_CREDIT_TYPE_Unknown type of loan_count_norm</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>100001</th>\n",
       "      <td>3</td>\n",
       "      <td>0.428571</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4</td>\n",
       "      <td>0.571429</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>7</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100002</th>\n",
       "      <td>2</td>\n",
       "      <td>0.250000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6</td>\n",
       "      <td>0.750000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>8</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100003</th>\n",
       "      <td>1</td>\n",
       "      <td>0.250000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3</td>\n",
       "      <td>0.750000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100004</th>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100005</th>\n",
       "      <td>2</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 46 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            bureau_CREDIT_ACTIVE_Active_count  \\\n",
       "SK_ID_CURR                                      \n",
       "100001                                      3   \n",
       "100002                                      2   \n",
       "100003                                      1   \n",
       "100004                                      0   \n",
       "100005                                      2   \n",
       "\n",
       "            bureau_CREDIT_ACTIVE_Active_count_norm  \\\n",
       "SK_ID_CURR                                           \n",
       "100001                                    0.428571   \n",
       "100002                                    0.250000   \n",
       "100003                                    0.250000   \n",
       "100004                                    0.000000   \n",
       "100005                                    0.666667   \n",
       "\n",
       "            bureau_CREDIT_ACTIVE_Bad debt_count  \\\n",
       "SK_ID_CURR                                        \n",
       "100001                                        0   \n",
       "100002                                        0   \n",
       "100003                                        0   \n",
       "100004                                        0   \n",
       "100005                                        0   \n",
       "\n",
       "            bureau_CREDIT_ACTIVE_Bad debt_count_norm  \\\n",
       "SK_ID_CURR                                             \n",
       "100001                                           0.0   \n",
       "100002                                           0.0   \n",
       "100003                                           0.0   \n",
       "100004                                           0.0   \n",
       "100005                                           0.0   \n",
       "\n",
       "            bureau_CREDIT_ACTIVE_Closed_count  \\\n",
       "SK_ID_CURR                                      \n",
       "100001                                      4   \n",
       "100002                                      6   \n",
       "100003                                      3   \n",
       "100004                                      2   \n",
       "100005                                      1   \n",
       "\n",
       "            bureau_CREDIT_ACTIVE_Closed_count_norm  \\\n",
       "SK_ID_CURR                                           \n",
       "100001                                    0.571429   \n",
       "100002                                    0.750000   \n",
       "100003                                    0.750000   \n",
       "100004                                    1.000000   \n",
       "100005                                    0.333333   \n",
       "\n",
       "            bureau_CREDIT_ACTIVE_Sold_count  \\\n",
       "SK_ID_CURR                                    \n",
       "100001                                    0   \n",
       "100002                                    0   \n",
       "100003                                    0   \n",
       "100004                                    0   \n",
       "100005                                    0   \n",
       "\n",
       "            bureau_CREDIT_ACTIVE_Sold_count_norm  \\\n",
       "SK_ID_CURR                                         \n",
       "100001                                       0.0   \n",
       "100002                                       0.0   \n",
       "100003                                       0.0   \n",
       "100004                                       0.0   \n",
       "100005                                       0.0   \n",
       "\n",
       "            bureau_CREDIT_CURRENCY_currency 1_count  \\\n",
       "SK_ID_CURR                                            \n",
       "100001                                            7   \n",
       "100002                                            8   \n",
       "100003                                            4   \n",
       "100004                                            2   \n",
       "100005                                            3   \n",
       "\n",
       "            bureau_CREDIT_CURRENCY_currency 1_count_norm  \\\n",
       "SK_ID_CURR                                                 \n",
       "100001                                               1.0   \n",
       "100002                                               1.0   \n",
       "100003                                               1.0   \n",
       "100004                                               1.0   \n",
       "100005                                               1.0   \n",
       "\n",
       "                                   ...                          \\\n",
       "SK_ID_CURR                         ...                           \n",
       "100001                             ...                           \n",
       "100002                             ...                           \n",
       "100003                             ...                           \n",
       "100004                             ...                           \n",
       "100005                             ...                           \n",
       "\n",
       "            bureau_CREDIT_TYPE_Microloan_count  \\\n",
       "SK_ID_CURR                                       \n",
       "100001                                       0   \n",
       "100002                                       0   \n",
       "100003                                       0   \n",
       "100004                                       0   \n",
       "100005                                       0   \n",
       "\n",
       "            bureau_CREDIT_TYPE_Microloan_count_norm  \\\n",
       "SK_ID_CURR                                            \n",
       "100001                                          0.0   \n",
       "100002                                          0.0   \n",
       "100003                                          0.0   \n",
       "100004                                          0.0   \n",
       "100005                                          0.0   \n",
       "\n",
       "            bureau_CREDIT_TYPE_Mobile operator loan_count  \\\n",
       "SK_ID_CURR                                                  \n",
       "100001                                                  0   \n",
       "100002                                                  0   \n",
       "100003                                                  0   \n",
       "100004                                                  0   \n",
       "100005                                                  0   \n",
       "\n",
       "            bureau_CREDIT_TYPE_Mobile operator loan_count_norm  \\\n",
       "SK_ID_CURR                                                       \n",
       "100001                                                    0.0    \n",
       "100002                                                    0.0    \n",
       "100003                                                    0.0    \n",
       "100004                                                    0.0    \n",
       "100005                                                    0.0    \n",
       "\n",
       "            bureau_CREDIT_TYPE_Mortgage_count  \\\n",
       "SK_ID_CURR                                      \n",
       "100001                                      0   \n",
       "100002                                      0   \n",
       "100003                                      0   \n",
       "100004                                      0   \n",
       "100005                                      0   \n",
       "\n",
       "            bureau_CREDIT_TYPE_Mortgage_count_norm  \\\n",
       "SK_ID_CURR                                           \n",
       "100001                                         0.0   \n",
       "100002                                         0.0   \n",
       "100003                                         0.0   \n",
       "100004                                         0.0   \n",
       "100005                                         0.0   \n",
       "\n",
       "            bureau_CREDIT_TYPE_Real estate loan_count  \\\n",
       "SK_ID_CURR                                              \n",
       "100001                                              0   \n",
       "100002                                              0   \n",
       "100003                                              0   \n",
       "100004                                              0   \n",
       "100005                                              0   \n",
       "\n",
       "            bureau_CREDIT_TYPE_Real estate loan_count_norm  \\\n",
       "SK_ID_CURR                                                   \n",
       "100001                                                 0.0   \n",
       "100002                                                 0.0   \n",
       "100003                                                 0.0   \n",
       "100004                                                 0.0   \n",
       "100005                                                 0.0   \n",
       "\n",
       "            bureau_CREDIT_TYPE_Unknown type of loan_count  \\\n",
       "SK_ID_CURR                                                  \n",
       "100001                                                  0   \n",
       "100002                                                  0   \n",
       "100003                                                  0   \n",
       "100004                                                  0   \n",
       "100005                                                  0   \n",
       "\n",
       "            bureau_CREDIT_TYPE_Unknown type of loan_count_norm  \n",
       "SK_ID_CURR                                                      \n",
       "100001                                                    0.0   \n",
       "100002                                                    0.0   \n",
       "100003                                                    0.0   \n",
       "100004                                                    0.0   \n",
       "100005                                                    0.0   \n",
       "\n",
       "[5 rows x 46 columns]"
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bureau_counts = count_categorical(bureau, group_var = 'SK_ID_CURR', df_name = 'bureau')\n",
    "bureau_counts.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "c4c80230ee098fa39f4446b7161f105cd18965bc"
   },
   "source": [
    "### Aggregated Stats of Bureau Dataframe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {
    "_uuid": "5f9d2491e244da4811f78f0340fd9d42127e9076",
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th>bureau_DAYS_CREDIT_mean</th>\n",
       "      <th>bureau_DAYS_CREDIT_max</th>\n",
       "      <th>bureau_DAYS_CREDIT_min</th>\n",
       "      <th>bureau_DAYS_CREDIT_sum</th>\n",
       "      <th>bureau_CREDIT_DAY_OVERDUE_mean</th>\n",
       "      <th>bureau_CREDIT_DAY_OVERDUE_max</th>\n",
       "      <th>bureau_CREDIT_DAY_OVERDUE_min</th>\n",
       "      <th>bureau_CREDIT_DAY_OVERDUE_sum</th>\n",
       "      <th>bureau_DAYS_CREDIT_ENDDATE_mean</th>\n",
       "      <th>...</th>\n",
       "      <th>bureau_AMT_CREDIT_SUM_OVERDUE_sum</th>\n",
       "      <th>bureau_DAYS_CREDIT_UPDATE_mean</th>\n",
       "      <th>bureau_DAYS_CREDIT_UPDATE_max</th>\n",
       "      <th>bureau_DAYS_CREDIT_UPDATE_min</th>\n",
       "      <th>bureau_DAYS_CREDIT_UPDATE_sum</th>\n",
       "      <th>bureau_AMT_ANNUITY_mean</th>\n",
       "      <th>bureau_AMT_ANNUITY_max</th>\n",
       "      <th>bureau_AMT_ANNUITY_min</th>\n",
       "      <th>bureau_AMT_ANNUITY_sum</th>\n",
       "      <th>bureau_counts</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>100001</td>\n",
       "      <td>-735.000000</td>\n",
       "      <td>-49</td>\n",
       "      <td>-1572</td>\n",
       "      <td>-5145</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>82.428571</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-93.142857</td>\n",
       "      <td>-6</td>\n",
       "      <td>-155</td>\n",
       "      <td>-652</td>\n",
       "      <td>3545.357143</td>\n",
       "      <td>10822.5</td>\n",
       "      <td>0.0</td>\n",
       "      <td>24817.5</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>100002</td>\n",
       "      <td>-874.000000</td>\n",
       "      <td>-103</td>\n",
       "      <td>-1437</td>\n",
       "      <td>-6992</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-349.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-499.875000</td>\n",
       "      <td>-7</td>\n",
       "      <td>-1185</td>\n",
       "      <td>-3999</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>100003</td>\n",
       "      <td>-1400.750000</td>\n",
       "      <td>-606</td>\n",
       "      <td>-2586</td>\n",
       "      <td>-5603</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-544.500000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-816.000000</td>\n",
       "      <td>-43</td>\n",
       "      <td>-2131</td>\n",
       "      <td>-3264</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>100004</td>\n",
       "      <td>-867.000000</td>\n",
       "      <td>-408</td>\n",
       "      <td>-1326</td>\n",
       "      <td>-1734</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-488.500000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-532.000000</td>\n",
       "      <td>-382</td>\n",
       "      <td>-682</td>\n",
       "      <td>-1064</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>100005</td>\n",
       "      <td>-190.666667</td>\n",
       "      <td>-62</td>\n",
       "      <td>-373</td>\n",
       "      <td>-572</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>439.333333</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-54.333333</td>\n",
       "      <td>-11</td>\n",
       "      <td>-121</td>\n",
       "      <td>-163</td>\n",
       "      <td>1420.500000</td>\n",
       "      <td>4261.5</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4261.5</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 50 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_CURR  bureau_DAYS_CREDIT_mean  bureau_DAYS_CREDIT_max  \\\n",
       "0      100001              -735.000000                     -49   \n",
       "1      100002              -874.000000                    -103   \n",
       "2      100003             -1400.750000                    -606   \n",
       "3      100004              -867.000000                    -408   \n",
       "4      100005              -190.666667                     -62   \n",
       "\n",
       "   bureau_DAYS_CREDIT_min  bureau_DAYS_CREDIT_sum  \\\n",
       "0                   -1572                   -5145   \n",
       "1                   -1437                   -6992   \n",
       "2                   -2586                   -5603   \n",
       "3                   -1326                   -1734   \n",
       "4                    -373                    -572   \n",
       "\n",
       "   bureau_CREDIT_DAY_OVERDUE_mean  bureau_CREDIT_DAY_OVERDUE_max  \\\n",
       "0                             0.0                              0   \n",
       "1                             0.0                              0   \n",
       "2                             0.0                              0   \n",
       "3                             0.0                              0   \n",
       "4                             0.0                              0   \n",
       "\n",
       "   bureau_CREDIT_DAY_OVERDUE_min  bureau_CREDIT_DAY_OVERDUE_sum  \\\n",
       "0                              0                              0   \n",
       "1                              0                              0   \n",
       "2                              0                              0   \n",
       "3                              0                              0   \n",
       "4                              0                              0   \n",
       "\n",
       "   bureau_DAYS_CREDIT_ENDDATE_mean      ...        \\\n",
       "0                        82.428571      ...         \n",
       "1                      -349.000000      ...         \n",
       "2                      -544.500000      ...         \n",
       "3                      -488.500000      ...         \n",
       "4                       439.333333      ...         \n",
       "\n",
       "   bureau_AMT_CREDIT_SUM_OVERDUE_sum  bureau_DAYS_CREDIT_UPDATE_mean  \\\n",
       "0                                0.0                      -93.142857   \n",
       "1                                0.0                     -499.875000   \n",
       "2                                0.0                     -816.000000   \n",
       "3                                0.0                     -532.000000   \n",
       "4                                0.0                      -54.333333   \n",
       "\n",
       "   bureau_DAYS_CREDIT_UPDATE_max  bureau_DAYS_CREDIT_UPDATE_min  \\\n",
       "0                             -6                           -155   \n",
       "1                             -7                          -1185   \n",
       "2                            -43                          -2131   \n",
       "3                           -382                           -682   \n",
       "4                            -11                           -121   \n",
       "\n",
       "   bureau_DAYS_CREDIT_UPDATE_sum  bureau_AMT_ANNUITY_mean  \\\n",
       "0                           -652              3545.357143   \n",
       "1                          -3999                 0.000000   \n",
       "2                          -3264                      NaN   \n",
       "3                          -1064                      NaN   \n",
       "4                           -163              1420.500000   \n",
       "\n",
       "   bureau_AMT_ANNUITY_max  bureau_AMT_ANNUITY_min  bureau_AMT_ANNUITY_sum  \\\n",
       "0                 10822.5                     0.0                 24817.5   \n",
       "1                     0.0                     0.0                     0.0   \n",
       "2                     NaN                     NaN                     0.0   \n",
       "3                     NaN                     NaN                     0.0   \n",
       "4                  4261.5                     0.0                  4261.5   \n",
       "\n",
       "   bureau_counts  \n",
       "0              7  \n",
       "1              8  \n",
       "2              4  \n",
       "3              2  \n",
       "4              3  \n",
       "\n",
       "[5 rows x 50 columns]"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bureau_agg = agg_numeric(bureau.drop(columns = ['SK_ID_BUREAU']), group_var = 'SK_ID_CURR', df_name = 'bureau')\n",
    "bureau_agg.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "3ce2211679505e646121cc0b2f8b217ab9718ba6"
   },
   "source": [
    "### Value counts of Bureau Balance dataframe by loan"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {
    "_uuid": "0e55cc0b41e66e77dfd19ac1b808edc4581a1206",
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>bureau_balance_STATUS_0_count</th>\n",
       "      <th>bureau_balance_STATUS_0_count_norm</th>\n",
       "      <th>bureau_balance_STATUS_1_count</th>\n",
       "      <th>bureau_balance_STATUS_1_count_norm</th>\n",
       "      <th>bureau_balance_STATUS_2_count</th>\n",
       "      <th>bureau_balance_STATUS_2_count_norm</th>\n",
       "      <th>bureau_balance_STATUS_3_count</th>\n",
       "      <th>bureau_balance_STATUS_3_count_norm</th>\n",
       "      <th>bureau_balance_STATUS_4_count</th>\n",
       "      <th>bureau_balance_STATUS_4_count_norm</th>\n",
       "      <th>bureau_balance_STATUS_5_count</th>\n",
       "      <th>bureau_balance_STATUS_5_count_norm</th>\n",
       "      <th>bureau_balance_STATUS_C_count</th>\n",
       "      <th>bureau_balance_STATUS_C_count_norm</th>\n",
       "      <th>bureau_balance_STATUS_X_count</th>\n",
       "      <th>bureau_balance_STATUS_X_count_norm</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_ID_BUREAU</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>5001709</th>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>86</td>\n",
       "      <td>0.886598</td>\n",
       "      <td>11</td>\n",
       "      <td>0.113402</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5001710</th>\n",
       "      <td>5</td>\n",
       "      <td>0.060241</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>48</td>\n",
       "      <td>0.578313</td>\n",
       "      <td>30</td>\n",
       "      <td>0.361446</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5001711</th>\n",
       "      <td>3</td>\n",
       "      <td>0.750000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1</td>\n",
       "      <td>0.250000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5001712</th>\n",
       "      <td>10</td>\n",
       "      <td>0.526316</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>9</td>\n",
       "      <td>0.473684</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5001713</th>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>22</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              bureau_balance_STATUS_0_count  \\\n",
       "SK_ID_BUREAU                                  \n",
       "5001709                                   0   \n",
       "5001710                                   5   \n",
       "5001711                                   3   \n",
       "5001712                                  10   \n",
       "5001713                                   0   \n",
       "\n",
       "              bureau_balance_STATUS_0_count_norm  \\\n",
       "SK_ID_BUREAU                                       \n",
       "5001709                                 0.000000   \n",
       "5001710                                 0.060241   \n",
       "5001711                                 0.750000   \n",
       "5001712                                 0.526316   \n",
       "5001713                                 0.000000   \n",
       "\n",
       "              bureau_balance_STATUS_1_count  \\\n",
       "SK_ID_BUREAU                                  \n",
       "5001709                                   0   \n",
       "5001710                                   0   \n",
       "5001711                                   0   \n",
       "5001712                                   0   \n",
       "5001713                                   0   \n",
       "\n",
       "              bureau_balance_STATUS_1_count_norm  \\\n",
       "SK_ID_BUREAU                                       \n",
       "5001709                                      0.0   \n",
       "5001710                                      0.0   \n",
       "5001711                                      0.0   \n",
       "5001712                                      0.0   \n",
       "5001713                                      0.0   \n",
       "\n",
       "              bureau_balance_STATUS_2_count  \\\n",
       "SK_ID_BUREAU                                  \n",
       "5001709                                   0   \n",
       "5001710                                   0   \n",
       "5001711                                   0   \n",
       "5001712                                   0   \n",
       "5001713                                   0   \n",
       "\n",
       "              bureau_balance_STATUS_2_count_norm  \\\n",
       "SK_ID_BUREAU                                       \n",
       "5001709                                      0.0   \n",
       "5001710                                      0.0   \n",
       "5001711                                      0.0   \n",
       "5001712                                      0.0   \n",
       "5001713                                      0.0   \n",
       "\n",
       "              bureau_balance_STATUS_3_count  \\\n",
       "SK_ID_BUREAU                                  \n",
       "5001709                                   0   \n",
       "5001710                                   0   \n",
       "5001711                                   0   \n",
       "5001712                                   0   \n",
       "5001713                                   0   \n",
       "\n",
       "              bureau_balance_STATUS_3_count_norm  \\\n",
       "SK_ID_BUREAU                                       \n",
       "5001709                                      0.0   \n",
       "5001710                                      0.0   \n",
       "5001711                                      0.0   \n",
       "5001712                                      0.0   \n",
       "5001713                                      0.0   \n",
       "\n",
       "              bureau_balance_STATUS_4_count  \\\n",
       "SK_ID_BUREAU                                  \n",
       "5001709                                   0   \n",
       "5001710                                   0   \n",
       "5001711                                   0   \n",
       "5001712                                   0   \n",
       "5001713                                   0   \n",
       "\n",
       "              bureau_balance_STATUS_4_count_norm  \\\n",
       "SK_ID_BUREAU                                       \n",
       "5001709                                      0.0   \n",
       "5001710                                      0.0   \n",
       "5001711                                      0.0   \n",
       "5001712                                      0.0   \n",
       "5001713                                      0.0   \n",
       "\n",
       "              bureau_balance_STATUS_5_count  \\\n",
       "SK_ID_BUREAU                                  \n",
       "5001709                                   0   \n",
       "5001710                                   0   \n",
       "5001711                                   0   \n",
       "5001712                                   0   \n",
       "5001713                                   0   \n",
       "\n",
       "              bureau_balance_STATUS_5_count_norm  \\\n",
       "SK_ID_BUREAU                                       \n",
       "5001709                                      0.0   \n",
       "5001710                                      0.0   \n",
       "5001711                                      0.0   \n",
       "5001712                                      0.0   \n",
       "5001713                                      0.0   \n",
       "\n",
       "              bureau_balance_STATUS_C_count  \\\n",
       "SK_ID_BUREAU                                  \n",
       "5001709                                  86   \n",
       "5001710                                  48   \n",
       "5001711                                   0   \n",
       "5001712                                   9   \n",
       "5001713                                   0   \n",
       "\n",
       "              bureau_balance_STATUS_C_count_norm  \\\n",
       "SK_ID_BUREAU                                       \n",
       "5001709                                 0.886598   \n",
       "5001710                                 0.578313   \n",
       "5001711                                 0.000000   \n",
       "5001712                                 0.473684   \n",
       "5001713                                 0.000000   \n",
       "\n",
       "              bureau_balance_STATUS_X_count  \\\n",
       "SK_ID_BUREAU                                  \n",
       "5001709                                  11   \n",
       "5001710                                  30   \n",
       "5001711                                   1   \n",
       "5001712                                   0   \n",
       "5001713                                  22   \n",
       "\n",
       "              bureau_balance_STATUS_X_count_norm  \n",
       "SK_ID_BUREAU                                      \n",
       "5001709                                 0.113402  \n",
       "5001710                                 0.361446  \n",
       "5001711                                 0.250000  \n",
       "5001712                                 0.000000  \n",
       "5001713                                 1.000000  "
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bureau_balance_counts = count_categorical(bureau_balance, group_var = 'SK_ID_BUREAU', df_name = 'bureau_balance')\n",
    "bureau_balance_counts.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "7c2378a4fc130bc7de954c454d50e26adda6f9f6"
   },
   "source": [
    "### Aggregated stats of Bureau Balance dataframe by loan"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {
    "_uuid": "eb25294df7f998058e15e6ae24438c772d4ae86c",
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_BUREAU</th>\n",
       "      <th>bureau_balance_MONTHS_BALANCE_mean</th>\n",
       "      <th>bureau_balance_MONTHS_BALANCE_max</th>\n",
       "      <th>bureau_balance_MONTHS_BALANCE_min</th>\n",
       "      <th>bureau_balance_MONTHS_BALANCE_sum</th>\n",
       "      <th>bureau_balance_counts</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5001709</td>\n",
       "      <td>-48.0</td>\n",
       "      <td>0</td>\n",
       "      <td>-96</td>\n",
       "      <td>-4656</td>\n",
       "      <td>97</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5001710</td>\n",
       "      <td>-41.0</td>\n",
       "      <td>0</td>\n",
       "      <td>-82</td>\n",
       "      <td>-3403</td>\n",
       "      <td>83</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>5001711</td>\n",
       "      <td>-1.5</td>\n",
       "      <td>0</td>\n",
       "      <td>-3</td>\n",
       "      <td>-6</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>5001712</td>\n",
       "      <td>-9.0</td>\n",
       "      <td>0</td>\n",
       "      <td>-18</td>\n",
       "      <td>-171</td>\n",
       "      <td>19</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5001713</td>\n",
       "      <td>-10.5</td>\n",
       "      <td>0</td>\n",
       "      <td>-21</td>\n",
       "      <td>-231</td>\n",
       "      <td>22</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_BUREAU  bureau_balance_MONTHS_BALANCE_mean  \\\n",
       "0       5001709                               -48.0   \n",
       "1       5001710                               -41.0   \n",
       "2       5001711                                -1.5   \n",
       "3       5001712                                -9.0   \n",
       "4       5001713                               -10.5   \n",
       "\n",
       "   bureau_balance_MONTHS_BALANCE_max  bureau_balance_MONTHS_BALANCE_min  \\\n",
       "0                                  0                                -96   \n",
       "1                                  0                                -82   \n",
       "2                                  0                                 -3   \n",
       "3                                  0                                -18   \n",
       "4                                  0                                -21   \n",
       "\n",
       "   bureau_balance_MONTHS_BALANCE_sum  bureau_balance_counts  \n",
       "0                              -4656                     97  \n",
       "1                              -3403                     83  \n",
       "2                                 -6                      4  \n",
       "3                               -171                     19  \n",
       "4                               -231                     22  "
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bureau_balance_agg = agg_numeric(bureau_balance, group_var = 'SK_ID_BUREAU', df_name = 'bureau_balance')\n",
    "bureau_balance_agg.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "f7343f605c4bc38c0c3af1b53bc258699e480dcf"
   },
   "source": [
    "### Aggregated Stats of Bureau Balance by Client"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {
    "_uuid": "9315370e38f25339af200a45809387f211464ce9",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Dataframe grouped by the loan\n",
    "bureau_by_loan = bureau_balance_agg.merge(bureau_balance_counts, right_index = True, left_on = 'SK_ID_BUREAU', how = 'outer')\n",
    "\n",
    "# Merge to include the SK_ID_CURR\n",
    "bureau_by_loan = bureau[['SK_ID_BUREAU', 'SK_ID_CURR']].merge(bureau_by_loan, on = 'SK_ID_BUREAU', how = 'left')\n",
    "\n",
    "# Aggregate the stats for each client\n",
    "bureau_balance_by_client = agg_numeric(bureau_by_loan.drop(columns = ['SK_ID_BUREAU']), group_var = 'SK_ID_CURR', df_name = 'client')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "1f8f616b645ba974517be24adb30b4af4cf1afe3"
   },
   "source": [
    "## Insert Computed Features into Training Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {
    "_uuid": "f61f183707d7ca8d363a344b073dd9122acfdcc5",
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Original Number of Features:  122\n"
     ]
    }
   ],
   "source": [
    "original_features = list(train.columns)\n",
    "print('Original Number of Features: ', len(original_features))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {
    "_uuid": "02f4310f7485d299c1588805ce24f3360afa6f42",
    "collapsed": true
   },
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'bureau_by_client' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-84-d07dff77ddf4>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m()\u001b[0m\n\u001b[0;32m      6\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      7\u001b[0m \u001b[1;31m# Merge with the monthly information grouped by client\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m----> 8\u001b[1;33m \u001b[0mtrain\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mtrain\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mmerge\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mbureau_by_client\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mon\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;34m'SK_ID_CURR'\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mhow\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;34m'left'\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m: name 'bureau_by_client' is not defined"
     ]
    }
   ],
   "source": [
    "# Merge with the value counts of bureau\n",
    "train = train.merge(bureau_counts, on = 'SK_ID_CURR', how = 'left')\n",
    "\n",
    "# Merge with the stats of bureau\n",
    "train = train.merge(bureau_agg, on = 'SK_ID_CURR', how = 'left')\n",
    "\n",
    "# Merge with the monthly information grouped by client\n",
    "train = train.merge(bureau_balance_by_client, on = 'SK_ID_CURR', how = 'left')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {
    "_uuid": "f672226eee3059311d476c0eb68946988f9f5712",
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of features using previous loans from other institutions data:  302\n"
     ]
    }
   ],
   "source": [
    "new_features = list(train.columns)\n",
    "print('Number of features using previous loans from other institutions data: ', len(new_features))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "e16d0c10cf2a0cc778db576a4ad233960c70a028"
   },
   "source": [
    "# Feature Engineering Outcomes\n",
    "\n",
    "After all that work, now we want to take a look at the variables we have created. We can look at the percentage of missing values, the correlations of variables with the target, and also the correlation of variables with the other variables. The correlations between variables can show if we have collinear varibles, that is, variables that are highly correlated with one another. Often, we want to remove one in a pair of collinear variables because having both variables would be redundant. We can also use the percentage of missing values to remove features with a substantial majority of values that are not present. __Feature selection__ will be an important focus going forward, because reducing the number of features can help the model learn during training and also generalize better to the testing data. The \"curse of dimensionality\" is the name given to the issues caused by having too many features (too high of a dimension). As the number of variables increases, the number of datapoints needed to learn the relationship between these variables and the target value increases exponentially. \n",
    "\n",
    "Feature selection is the process of removing variables to help our model to learn and generalize better to the testing set. The objective is to remove useless/redundant variables while preserving those that are useful. There are a number of tools we can use for this process, but in this notebook we will stick to removing columns with a high percentage of missing values and variables that have a high correlation with one another. Later we can look at using the feature importances returned from models such as the `Gradient Boosting Machine` or `Random Forest` to perform feature selection."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "ab576d2c245270ca843f825afab02583c091897d"
   },
   "source": [
    "## Missing Values\n",
    "\n",
    "An important consideration is the missing values in the dataframe. Columns with too many missing values might have to be dropped. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {
    "_uuid": "eef874f5edd28e7449e198264987b93566a5227a",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Function to calculate missing values by column# Funct \n",
    "def missing_values_table(df):\n",
    "        # Total missing values\n",
    "        mis_val = df.isnull().sum()\n",
    "        \n",
    "        # Percentage of missing values\n",
    "        mis_val_percent = 100 * df.isnull().sum() / len(df)\n",
    "        \n",
    "        # Make a table with the results\n",
    "        mis_val_table = pd.concat([mis_val, mis_val_percent], axis=1)\n",
    "        \n",
    "        # Rename the columns\n",
    "        mis_val_table_ren_columns = mis_val_table.rename(\n",
    "        columns = {0 : 'Missing Values', 1 : '% of Total Values'})\n",
    "        \n",
    "        # Sort the table by percentage of missing descending\n",
    "        mis_val_table_ren_columns = mis_val_table_ren_columns[\n",
    "            mis_val_table_ren_columns.iloc[:,1] != 0].sort_values(\n",
    "        '% of Total Values', ascending=False).round(1)\n",
    "        \n",
    "        # Print some summary information\n",
    "        print (\"Your selected dataframe has \" + str(df.shape[1]) + \" columns.\\n\"      \n",
    "            \"There are \" + str(mis_val_table_ren_columns.shape[0]) +\n",
    "              \" columns that have missing values.\")\n",
    "        \n",
    "        # Return the dataframe with missing information\n",
    "        return mis_val_table_ren_columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {
    "_uuid": "dba1e683347bbd9b07a162675a4f79e08c28db0a",
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Your selected dataframe has 302 columns.\n",
      "There are 247 columns that have missing values.\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Missing Values</th>\n",
       "      <th>% of Total Values</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>bureau_AMT_ANNUITY_mean</th>\n",
       "      <td>227502</td>\n",
       "      <td>74.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>bureau_AMT_ANNUITY_min</th>\n",
       "      <td>227502</td>\n",
       "      <td>74.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>bureau_AMT_ANNUITY_max</th>\n",
       "      <td>227502</td>\n",
       "      <td>74.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>client_bureau_balance_STATUS_C_count_norm_max</th>\n",
       "      <td>215280</td>\n",
       "      <td>70.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>client_bureau_balance_STATUS_4_count_min</th>\n",
       "      <td>215280</td>\n",
       "      <td>70.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>client_bureau_balance_STATUS_2_count_norm_mean</th>\n",
       "      <td>215280</td>\n",
       "      <td>70.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>client_bureau_balance_STATUS_2_count_norm_max</th>\n",
       "      <td>215280</td>\n",
       "      <td>70.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>client_bureau_balance_STATUS_2_count_norm_min</th>\n",
       "      <td>215280</td>\n",
       "      <td>70.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>client_bureau_balance_STATUS_3_count_mean</th>\n",
       "      <td>215280</td>\n",
       "      <td>70.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>client_bureau_balance_STATUS_3_count_max</th>\n",
       "      <td>215280</td>\n",
       "      <td>70.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                Missing Values  \\\n",
       "bureau_AMT_ANNUITY_mean                                 227502   \n",
       "bureau_AMT_ANNUITY_min                                  227502   \n",
       "bureau_AMT_ANNUITY_max                                  227502   \n",
       "client_bureau_balance_STATUS_C_count_norm_max           215280   \n",
       "client_bureau_balance_STATUS_4_count_min                215280   \n",
       "client_bureau_balance_STATUS_2_count_norm_mean          215280   \n",
       "client_bureau_balance_STATUS_2_count_norm_max           215280   \n",
       "client_bureau_balance_STATUS_2_count_norm_min           215280   \n",
       "client_bureau_balance_STATUS_3_count_mean               215280   \n",
       "client_bureau_balance_STATUS_3_count_max                215280   \n",
       "\n",
       "                                                % of Total Values  \n",
       "bureau_AMT_ANNUITY_mean                                      74.0  \n",
       "bureau_AMT_ANNUITY_min                                       74.0  \n",
       "bureau_AMT_ANNUITY_max                                       74.0  \n",
       "client_bureau_balance_STATUS_C_count_norm_max                70.0  \n",
       "client_bureau_balance_STATUS_4_count_min                     70.0  \n",
       "client_bureau_balance_STATUS_2_count_norm_mean               70.0  \n",
       "client_bureau_balance_STATUS_2_count_norm_max                70.0  \n",
       "client_bureau_balance_STATUS_2_count_norm_min                70.0  \n",
       "client_bureau_balance_STATUS_3_count_mean                    70.0  \n",
       "client_bureau_balance_STATUS_3_count_max                     70.0  "
      ]
     },
     "execution_count": 88,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "missing_train = missing_values_table(train)\n",
    "missing_train.head(10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "f0defbc704277c67350cf2064b511800a3cdbad2"
   },
   "source": [
    "We see there are a number of columns with a high percentage of missing values. There is no well-established threshold for removing missing values, and the best course of action depends on the problem. Here, to reduce the number of features, we will remove any columns in either the training or the testing data that have greater than 90% missing values."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {
    "_uuid": "5739842767f1a58e9c33a6683d5b0f32c97b69b3",
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 89,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "missing_train_vars = list(missing_train.index[missing_train['% of Total Values'] > 90])\n",
    "len(missing_train_vars)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "055b96cb462b6876d6ae7f0e0c348426ddfc5cca"
   },
   "source": [
    "Before we remove the missing values, we will find the missing value percentages in the testing data. We'll then remove any columns with greater than 90% missing values in either the training or testing data.\n",
    "Let's now read in the testing data, perform the same operations, and look at the missing values in the testing data. We already have calculated all the counts and aggregation statistics, so we only need to merge the testing data with the appropriate data. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "dc3921576a8fb644b36f5119f169b88d30a51c01"
   },
   "source": [
    "## Calculate Information for Testing Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {
    "_uuid": "48be68ba3403d8e3f208b436219003b1f6097fad",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Read in the test dataframe\n",
    "test = pd.read_csv('../input/application_test.csv')\n",
    "\n",
    "# Merge with the value counts of bureau\n",
    "test = test.merge(bureau_cats, on = 'SK_ID_CURR', how = 'left')\n",
    "\n",
    "# Merge with the stats of bureau\n",
    "test = test.merge(bureau_agg, on = 'SK_ID_CURR', how = 'left')\n",
    "\n",
    "# Merge with the value counts of bureau balance\n",
    "test = test.merge(bureau_balance_by_client, on = 'SK_ID_CURR', how = 'left')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {
    "_uuid": "cbbd8b82cf6e00e14a8679d3fb1cc2a63a7fcf61",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Fill the missing counts with 0\n",
    "test['bureau_counts'] = test['bureau_counts'].fillna(0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {
    "_uuid": "0ee6595328bd2eab6a89b542475500aaaa4b3947",
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape of Testing Data:  (48744, 301)\n"
     ]
    }
   ],
   "source": [
    "print('Shape of Testing Data: ', test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "518da486b84f28f7ca9c1b366fdb2719f0e19f38"
   },
   "source": [
    "We need to align the testing and training dataframes, which means matching up the columns so they have the exact same columns. This shouldn't be an issue here, but when we one-hot encode variables, we need to align the dataframes to make sure they have the same columns."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {
    "_uuid": "39d6e6a2b8af1bb15df0f98ff5eb601c2e828099",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train_labels = train['TARGET']\n",
    "\n",
    "# Align the dataframes, this will remove the 'TARGET' column\n",
    "train, test = train.align(test, join = 'inner', axis = 1)\n",
    "\n",
    "train['TARGET'] = train_labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {
    "_uuid": "f3d3fcc7cf9a68208a6d6303565883cd58e7c531",
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Data Shape:  (307511, 302)\n",
      "Testing Data Shape:  (48744, 301)\n"
     ]
    }
   ],
   "source": [
    "print('Training Data Shape: ', train.shape)\n",
    "print('Testing Data Shape: ', test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "77cd0cbc96158d555ca156b93906497364cbdabd"
   },
   "source": [
    "The dataframes now have the same columns (with the exception of the `TARGET` column in the training data). This means we can use them in a machine learning model which needs to see the same columns in both the training and testing dataframes.\n",
    "\n",
    "Let's now look at the percentage of missing values in the testing data so we can figure out the columns that should be dropped."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {
    "_uuid": "ff8e3932b44097ae679ea7bb9b0464a077e07727",
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Your selected dataframe has 301 columns.\n",
      "There are 243 columns that have missing values.\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Missing Values</th>\n",
       "      <th>% of Total Values</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>COMMONAREA_MEDI</th>\n",
       "      <td>33495</td>\n",
       "      <td>68.7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>COMMONAREA_MODE</th>\n",
       "      <td>33495</td>\n",
       "      <td>68.7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>COMMONAREA_AVG</th>\n",
       "      <td>33495</td>\n",
       "      <td>68.7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NONLIVINGAPARTMENTS_MEDI</th>\n",
       "      <td>33347</td>\n",
       "      <td>68.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NONLIVINGAPARTMENTS_MODE</th>\n",
       "      <td>33347</td>\n",
       "      <td>68.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NONLIVINGAPARTMENTS_AVG</th>\n",
       "      <td>33347</td>\n",
       "      <td>68.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FONDKAPREMONT_MODE</th>\n",
       "      <td>32797</td>\n",
       "      <td>67.3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LIVINGAPARTMENTS_MEDI</th>\n",
       "      <td>32780</td>\n",
       "      <td>67.2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LIVINGAPARTMENTS_MODE</th>\n",
       "      <td>32780</td>\n",
       "      <td>67.2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LIVINGAPARTMENTS_AVG</th>\n",
       "      <td>32780</td>\n",
       "      <td>67.2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                          Missing Values  % of Total Values\n",
       "COMMONAREA_MEDI                    33495               68.7\n",
       "COMMONAREA_MODE                    33495               68.7\n",
       "COMMONAREA_AVG                     33495               68.7\n",
       "NONLIVINGAPARTMENTS_MEDI           33347               68.4\n",
       "NONLIVINGAPARTMENTS_MODE           33347               68.4\n",
       "NONLIVINGAPARTMENTS_AVG            33347               68.4\n",
       "FONDKAPREMONT_MODE                 32797               67.3\n",
       "LIVINGAPARTMENTS_MEDI              32780               67.2\n",
       "LIVINGAPARTMENTS_MODE              32780               67.2\n",
       "LIVINGAPARTMENTS_AVG               32780               67.2"
      ]
     },
     "execution_count": 97,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "missing_test = missing_values_table(test)\n",
    "missing_test.head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {
    "_uuid": "1ecf51b6455fdef32ea7b50cdaf05eed211b15e7",
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 98,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "missing_test_vars = list(missing_test.index[missing_test['% of Total Values'] > 90])\n",
    "len(missing_test_vars)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {
    "_uuid": "caa333e30cf19898f4b902d64fbd790659c8f035",
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There are 0 columns with more than 90% missing in either the training or testing data.\n"
     ]
    }
   ],
   "source": [
    "missing_columns = list(set(missing_test_vars + missing_train_vars))\n",
    "print('There are %d columns with more than 90%% missing in either the training or testing data.' % len(missing_columns))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {
    "_uuid": "3ae7c69dc4112159c5e1356b06821ba169ac595d",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Drop the missing columns\n",
    "train = train.drop(columns = missing_columns)\n",
    "test = test.drop(columns = missing_columns)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We ended up removing no columns in this round because there are no columns with more than 90% missing values. We might have to apply another feature selection method to reduce the dimensionality. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "91197da37fcd84538520f03b50e0b060a19cc71d"
   },
   "source": [
    "At this point we will save both the training and testing data. I encourage anyone to try different percentages for dropping the missing columns and compare the outcomes. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {
    "_uuid": "a7d15638e419b84bf164b20e5195db37885e3e0d",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train.to_csv('train_bureau_raw.csv', index = False)\n",
    "test.to_csv('test_bureau_raw.csv', index = False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "9425c558df0c882cffe4fc07da41394a6fb3b422"
   },
   "source": [
    "## Correlations\n",
    "\n",
    "First let's look at the correlations of the variables with the target. We can see in any of the variables we created have a greater correlation than those already present in the training data (from `application`). "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {
    "_uuid": "87c47e087eaacb846572ffdd58c22ca1f0ce0010",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Calculate all correlations in dataframe\n",
    "corrs = train.corr()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {
    "_uuid": "75c757ae81ff84d9e3453ca7c74ce59326e24b84",
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>TARGET</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>TARGET</th>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>bureau_DAYS_CREDIT_mean</th>\n",
       "      <td>0.089729</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>client_bureau_balance_MONTHS_BALANCE_min_mean</th>\n",
       "      <td>0.089038</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_BIRTH</th>\n",
       "      <td>0.078239</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>bureau_CREDIT_ACTIVE_Active_count_norm</th>\n",
       "      <td>0.077356</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>client_bureau_balance_MONTHS_BALANCE_mean_mean</th>\n",
       "      <td>0.076424</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>bureau_DAYS_CREDIT_min</th>\n",
       "      <td>0.075248</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>client_bureau_balance_MONTHS_BALANCE_min_min</th>\n",
       "      <td>0.073225</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>client_bureau_balance_MONTHS_BALANCE_sum_mean</th>\n",
       "      <td>0.072606</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>bureau_DAYS_CREDIT_UPDATE_mean</th>\n",
       "      <td>0.068927</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                  TARGET\n",
       "TARGET                                          1.000000\n",
       "bureau_DAYS_CREDIT_mean                         0.089729\n",
       "client_bureau_balance_MONTHS_BALANCE_min_mean   0.089038\n",
       "DAYS_BIRTH                                      0.078239\n",
       "bureau_CREDIT_ACTIVE_Active_count_norm          0.077356\n",
       "client_bureau_balance_MONTHS_BALANCE_mean_mean  0.076424\n",
       "bureau_DAYS_CREDIT_min                          0.075248\n",
       "client_bureau_balance_MONTHS_BALANCE_min_min    0.073225\n",
       "client_bureau_balance_MONTHS_BALANCE_sum_mean   0.072606\n",
       "bureau_DAYS_CREDIT_UPDATE_mean                  0.068927"
      ]
     },
     "execution_count": 104,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "corrs = corrs.sort_values('TARGET', ascending = False)\n",
    "\n",
    "# Ten most positive correlations\n",
    "pd.DataFrame(corrs['TARGET'].head(10))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {
    "_uuid": "6774bdef9c97bc58191101414a0a067561f12160",
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>TARGET</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>client_bureau_balance_counts_min</th>\n",
       "      <td>-0.048224</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>client_bureau_balance_STATUS_C_count_norm_mean</th>\n",
       "      <td>-0.055936</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>client_bureau_balance_STATUS_C_count_max</th>\n",
       "      <td>-0.061083</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>client_bureau_balance_STATUS_C_count_mean</th>\n",
       "      <td>-0.062954</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>client_bureau_balance_counts_max</th>\n",
       "      <td>-0.068792</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>bureau_CREDIT_ACTIVE_Closed_count_norm</th>\n",
       "      <td>-0.079369</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>client_bureau_balance_counts_mean</th>\n",
       "      <td>-0.080193</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>EXT_SOURCE_1</th>\n",
       "      <td>-0.155317</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>EXT_SOURCE_2</th>\n",
       "      <td>-0.160472</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>EXT_SOURCE_3</th>\n",
       "      <td>-0.178919</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                  TARGET\n",
       "client_bureau_balance_counts_min               -0.048224\n",
       "client_bureau_balance_STATUS_C_count_norm_mean -0.055936\n",
       "client_bureau_balance_STATUS_C_count_max       -0.061083\n",
       "client_bureau_balance_STATUS_C_count_mean      -0.062954\n",
       "client_bureau_balance_counts_max               -0.068792\n",
       "bureau_CREDIT_ACTIVE_Closed_count_norm         -0.079369\n",
       "client_bureau_balance_counts_mean              -0.080193\n",
       "EXT_SOURCE_1                                   -0.155317\n",
       "EXT_SOURCE_2                                   -0.160472\n",
       "EXT_SOURCE_3                                   -0.178919"
      ]
     },
     "execution_count": 105,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Ten most negative correlations\n",
    "pd.DataFrame(corrs['TARGET'].dropna().tail(10))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "0d98ec8d33f097aa3702933cfb64d82bfbd62232"
   },
   "source": [
    "The highest correlated variable with the target (other than the `TARGET` which of course has a correlation of 1), is a variable we created. However, just because the variable is correlated does not mean that it will be useful, and we have to remember that if we generate hundreds of new variables, some are going to be correlated with the target simply because of random noise. \n",
    "\n",
    "Viewing the correlations skeptically, it does appear that several of the newly created variables may be useful. To assess the \"usefulness\" of variables, we will look at the feature importances returned by the model. For curiousity's sake (and because we already wrote the function) we can make a kde plot of two of the newly created variables."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {
    "_uuid": "c92a7810a7e0e6fd5d92e1be50749e427fae7c11",
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The correlation between client_bureau_balance_counts_mean and the TARGET is -0.0802\n",
      "Median value for loan that was not repaid = 19.3333\n",
      "Median value for loan that was repaid =     25.1429\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "kde_target(var_name='client_bureau_balance_counts_mean', df=train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "b8c47b227c6c4130ea9d3454a789b82c422d20a5"
   },
   "source": [
    "This variable represents the average number of monthly records per loan for each client. For example, if a client had three previous loans with 3, 4, and 5 records in the monthly data, the value of this variable for them would be 4. Based on the distribution, clients with a greater number of average monthly records per loan were more likely to repay their loans with Home Credit. Let's not read too much into this value, but it could indicate that clients who have had more previous credit history are generally more likely to repay a loan."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {
    "_uuid": "17e1cb7c492b1c122556ac16472fbbdee5a1030a",
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The correlation between bureau_CREDIT_ACTIVE_Active_count_norm and the TARGET is 0.0774\n",
      "Median value for loan that was not repaid = 0.5000\n",
      "Median value for loan that was repaid =     0.3636\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "kde_target(var_name='bureau_CREDIT_ACTIVE_Active_count_norm', df=train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "dd71a301ca0fe799b5f36ddd07cf3224c03e1cd9"
   },
   "source": [
    "Well this distribution is all over the place. This variable represents the number of previous loans with a `CREDIT_ACTIVE` value of `Active` divided by the total number of previous loans for a client. The correlation here is so weak that I do not think we should draw any conclusions! "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "e6ed8ce92d28415771fe4e4cc872117a18357181"
   },
   "source": [
    "### Collinear Variables\n",
    "\n",
    "We can calculate not only the correlations of the variables with the target, but also the correlation of each variable with every other variable. This will allow us to see if there are highly collinear variables that should perhaps be removed from the data. \n",
    "\n",
    "Let's look for any variables that have a greather than 0.8 correlation with other variables."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "metadata": {
    "_uuid": "7330e7311864a70afa75b7e7c33ad4b3d77df040",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Set the threshold\n",
    "threshold = 0.8\n",
    "\n",
    "# Empty dictionary to hold correlated variables\n",
    "above_threshold_vars = {}\n",
    "\n",
    "# For each column, record the variables that are above the threshold\n",
    "for col in corrs:\n",
    "    above_threshold_vars[col] = list(corrs.index[corrs[col] > threshold])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "8dcfa3892d84efa2e0db448b9542c9effad01e85"
   },
   "source": [
    "For each of these pairs of highly correlated variables, we only want to remove one of the variables. The following code creates a set of variables to remove by only adding one of each pair. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {
    "_uuid": "0134543670e895436769c64a58bbdee7eb7a255f",
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of columns to remove:  104\n"
     ]
    }
   ],
   "source": [
    "# Track columns to remove and columns already examined\n",
    "cols_to_remove = []\n",
    "cols_seen = []\n",
    "cols_to_remove_pair = []\n",
    "\n",
    "# Iterate through columns and correlated columns\n",
    "for key, value in above_threshold_vars.items():\n",
    "    # Keep track of columns already examined\n",
    "    cols_seen.append(key)\n",
    "    for x in value:\n",
    "        if x == key:\n",
    "            next\n",
    "        else:\n",
    "            # Only want to remove one in a pair\n",
    "            if x not in cols_seen:\n",
    "                cols_to_remove.append(x)\n",
    "                cols_to_remove_pair.append(key)\n",
    "            \n",
    "cols_to_remove = list(set(cols_to_remove))\n",
    "print('Number of columns to remove: ', len(cols_to_remove))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "be37ec11c061291e143c01102703103c0915ffcb"
   },
   "source": [
    "We can remove these columns from both the training and the testing datasets. We will have to compare performance after removing these variables with performance keeping these variables (the raw csv files we saved earlier)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {
    "_uuid": "0fedcbce9bce99809bf7d4559b6ad93e8ecca8ea",
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Corrs Removed Shape:  (307511, 198)\n",
      "Testing Corrs Removed Shape:  (48744, 197)\n"
     ]
    }
   ],
   "source": [
    "train_corrs_removed = train.drop(columns = cols_to_remove)\n",
    "test_corrs_removed = test.drop(columns = cols_to_remove)\n",
    "\n",
    "print('Training Corrs Removed Shape: ', train_corrs_removed.shape)\n",
    "print('Testing Corrs Removed Shape: ', test_corrs_removed.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "_uuid": "2f4443a62f6e8987c0be2e3162589444ecd7d275",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train_corrs_removed.to_csv('train_bureau_corrs_removed.csv', index = False)\n",
    "test_corrs_removed.to_csv('test_bureau_corrs_removed.csv', index = False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "be48572e88f64f00b4221da7782ac834444c96b0"
   },
   "source": [
    "# Modeling \n",
    "\n",
    "To actually test the performance of these new datasets, we will try using them for machine learning! Here we will use a function I developed in another notebook to compare the features (the raw version with the highly correlated variables removed). We can run this kind of like an experiment, and the control will be the performance of just the `application` data in this function when submitted to the competition. I've already recorded that performance, \n",
    "so we can list out our control and our two test conditions:\n",
    "\n",
    "__For all datasets, use the model shown below (with the exact hyperparameters).__\n",
    "\n",
    "* control: only the data in the `application` files. \n",
    "* test one: the data in the `application` files with all of the data recorded from the `bureau` and `bureau_balance` files\n",
    "* test two: the data in the `application` files with all of the data recorded from the `bureau` and `bureau_balance` files with highly correlated variables removed. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {
    "_uuid": "58aac4a109bde812ca8b75b1c9271550dcf2d397",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import lightgbm as lgb\n",
    "\n",
    "from sklearn.model_selection import KFold\n",
    "from sklearn.metrics import roc_auc_score\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {
    "_uuid": "ae2c8a505885482595f027e6139b55966a88861e",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def model(features, test_features, encoding = 'ohe', n_folds = 5):\n",
    "    \n",
    "    \"\"\"Train and test a light gradient boosting model using\n",
    "    cross validation. \n",
    "    \n",
    "    Args:\n",
    "        features (pd.DataFrame): dataframe of training features to use \n",
    "        for training a model. Must include the TARGET column.\n",
    "        test_features (pd.DataFrame): dataframe of testing features to use\n",
    "        for making predictions with the model. \n",
    "        encoding (str, default = 'ohe'): method for encoding categorical \n",
    "        variables. Either 'ohe' for one-hot encoding or 'le' for integer label encoding\n",
    "        n_folds (int, default = 5): number of folds to use for cross validation\n",
    "        \n",
    "    Returns:\n",
    "        submission (pd.DataFrame): dataframe with `SK_ID_CURR` and `TARGET` probabilities\n",
    "        predicted by the model.\n",
    "        feature_importances (pd.DataFrame): dataframe with the feature importances\n",
    "        from the model.\n",
    "        valid_metrics (pd.DataFrame): dataframe with validation metrics (ROC AUC) for \n",
    "        each fold and overall.\n",
    "        \n",
    "    \"\"\"\n",
    "    \n",
    "    # Extract the ids\n",
    "    train_ids = features['SK_ID_CURR']\n",
    "    test_ids = test_features['SK_ID_CURR']\n",
    "    \n",
    "    # Extract the labels for training\n",
    "    labels = features['TARGET']\n",
    "    \n",
    "    # Remove the ids and target\n",
    "    features = features.drop(columns = ['SK_ID_CURR', 'TARGET'])\n",
    "    test_features = test_features.drop(columns = ['SK_ID_CURR'])\n",
    "    \n",
    "    \n",
    "    # One Hot Encoding\n",
    "    if encoding == 'ohe':\n",
    "        features = pd.get_dummies(features)\n",
    "        test_features = pd.get_dummies(test_features)\n",
    "        \n",
    "        # Align the dataframes by the columns\n",
    "        features, test_features = features.align(test_features, join = 'inner', axis = 1)\n",
    "        \n",
    "        # No categorical indices to record\n",
    "        cat_indices = 'auto'\n",
    "    \n",
    "    # Integer label encoding\n",
    "    elif encoding == 'le':\n",
    "        \n",
    "        # Create a label encoder\n",
    "        label_encoder = LabelEncoder()\n",
    "        \n",
    "        # List for storing categorical indices\n",
    "        cat_indices = []\n",
    "        \n",
    "        # Iterate through each column\n",
    "        for i, col in enumerate(features):\n",
    "            if features[col].dtype == 'object':\n",
    "                # Map the categorical features to integers\n",
    "                features[col] = label_encoder.fit_transform(np.array(features[col].astype(str)).reshape((-1,)))\n",
    "                test_features[col] = label_encoder.transform(np.array(test_features[col].astype(str)).reshape((-1,)))\n",
    "\n",
    "                # Record the categorical indices\n",
    "                cat_indices.append(i)\n",
    "    \n",
    "    # Catch error if label encoding scheme is not valid\n",
    "    else:\n",
    "        raise ValueError(\"Encoding must be either 'ohe' or 'le'\")\n",
    "        \n",
    "    print('Training Data Shape: ', features.shape)\n",
    "    print('Testing Data Shape: ', test_features.shape)\n",
    "    \n",
    "    # Extract feature names\n",
    "    feature_names = list(features.columns)\n",
    "    \n",
    "    # Convert to np arrays\n",
    "    features = np.array(features)\n",
    "    test_features = np.array(test_features)\n",
    "    \n",
    "    # Create the kfold object\n",
    "    k_fold = KFold(n_splits = n_folds, shuffle = False, random_state = 50)\n",
    "    \n",
    "    # Empty array for feature importances\n",
    "    feature_importance_values = np.zeros(len(feature_names))\n",
    "    \n",
    "    # Empty array for test predictions\n",
    "    test_predictions = np.zeros(test_features.shape[0])\n",
    "    \n",
    "    # Empty array for out of fold validation predictions\n",
    "    out_of_fold = np.zeros(features.shape[0])\n",
    "    \n",
    "    # Lists for recording validation and training scores\n",
    "    valid_scores = []\n",
    "    train_scores = []\n",
    "    \n",
    "    # Iterate through each fold\n",
    "    for train_indices, valid_indices in k_fold.split(features):\n",
    "        \n",
    "        # Training data for the fold\n",
    "        train_features, train_labels = features[train_indices], labels[train_indices]\n",
    "        # Validation data for the fold\n",
    "        valid_features, valid_labels = features[valid_indices], labels[valid_indices]\n",
    "        \n",
    "        # Create the model\n",
    "        model = lgb.LGBMClassifier(n_estimators=10000, objective = 'binary', \n",
    "                                   class_weight = 'balanced', learning_rate = 0.05, \n",
    "                                   reg_alpha = 0.1, reg_lambda = 0.1, \n",
    "                                   subsample = 0.8, n_jobs = -1, random_state = 50)\n",
    "        \n",
    "        # Train the model\n",
    "        model.fit(train_features, train_labels, eval_metric = 'auc',\n",
    "                  eval_set = [(valid_features, valid_labels), (train_features, train_labels)],\n",
    "                  eval_names = ['valid', 'train'], categorical_feature = cat_indices,\n",
    "                  early_stopping_rounds = 100, verbose = 200)\n",
    "        \n",
    "        # Record the best iteration\n",
    "        best_iteration = model.best_iteration_\n",
    "        \n",
    "        # Record the feature importances\n",
    "        feature_importance_values += model.feature_importances_ / k_fold.n_splits\n",
    "        \n",
    "        # Make predictions\n",
    "        test_predictions += model.predict_proba(test_features, num_iteration = best_iteration)[:, 1] / k_fold.n_splits\n",
    "        \n",
    "        # Record the out of fold predictions\n",
    "        out_of_fold[valid_indices] = model.predict_proba(valid_features, num_iteration = best_iteration)[:, 1]\n",
    "        \n",
    "        # Record the best score\n",
    "        valid_score = model.best_score_['valid']['auc']\n",
    "        train_score = model.best_score_['train']['auc']\n",
    "        \n",
    "        valid_scores.append(valid_score)\n",
    "        train_scores.append(train_score)\n",
    "        \n",
    "        # Clean up memory\n",
    "        gc.enable()\n",
    "        del model, train_features, valid_features\n",
    "        gc.collect()\n",
    "        \n",
    "    # Make the submission dataframe\n",
    "    submission = pd.DataFrame({'SK_ID_CURR': test_ids, 'TARGET': test_predictions})\n",
    "    \n",
    "    # Make the feature importance dataframe\n",
    "    feature_importances = pd.DataFrame({'feature': feature_names, 'importance': feature_importance_values})\n",
    "    \n",
    "    # Overall validation score\n",
    "    valid_auc = roc_auc_score(labels, out_of_fold)\n",
    "    \n",
    "    # Add the overall scores to the metrics\n",
    "    valid_scores.append(valid_auc)\n",
    "    train_scores.append(np.mean(train_scores))\n",
    "    \n",
    "    # Needed for creating dataframe of validation scores\n",
    "    fold_names = list(range(n_folds))\n",
    "    fold_names.append('overall')\n",
    "    \n",
    "    # Dataframe of validation scores\n",
    "    metrics = pd.DataFrame({'fold': fold_names,\n",
    "                            'train': train_scores,\n",
    "                            'valid': valid_scores}) \n",
    "    \n",
    "    return submission, feature_importances, metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {
    "_uuid": "6b39fe9584e0308b7e940c11e6de04b262d123c7",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def plot_feature_importances(df):\n",
    "    \"\"\"\n",
    "    Plot importances returned by a model. This can work with any measure of\n",
    "    feature importance provided that higher importance is better. \n",
    "    \n",
    "    Args:\n",
    "        df (dataframe): feature importances. Must have the features in a column\n",
    "        called `features` and the importances in a column called `importance\n",
    "        \n",
    "    Returns:\n",
    "        shows a plot of the 15 most importance features\n",
    "        \n",
    "        df (dataframe): feature importances sorted by importance (highest to lowest) \n",
    "        with a column for normalized importance\n",
    "        \"\"\"\n",
    "    \n",
    "    # Sort features according to importance\n",
    "    df = df.sort_values('importance', ascending = False).reset_index()\n",
    "    \n",
    "    # Normalize the feature importances to add up to one\n",
    "    df['importance_normalized'] = df['importance'] / df['importance'].sum()\n",
    "\n",
    "    # Make a horizontal bar chart of feature importances\n",
    "    plt.figure(figsize = (10, 6))\n",
    "    ax = plt.subplot()\n",
    "    \n",
    "    # Need to reverse the index to plot most important on top\n",
    "    ax.barh(list(reversed(list(df.index[:15]))), \n",
    "            df['importance_normalized'].head(15), \n",
    "            align = 'center', edgecolor = 'k')\n",
    "    \n",
    "    # Set the yticks and labels\n",
    "    ax.set_yticks(list(reversed(list(df.index[:15]))))\n",
    "    ax.set_yticklabels(df['feature'].head(15))\n",
    "    \n",
    "    # Plot labeling\n",
    "    plt.xlabel('Normalized Importance'); plt.title('Feature Importances')\n",
    "    plt.show()\n",
    "    \n",
    "    return df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "52cc81edc0b29905d2325f201f7902e96c321e32"
   },
   "source": [
    "### Control\n",
    "\n",
    "The first step in any experiment is establishing a control. For this we will use the function defined above (that implements a Gradient Boosting Machine model) and the single main data source (`application`). "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {
    "_uuid": "7b9f0e5faa3de9403f87032fe3fc339cb732ad17",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train_control = pd.read_csv('../input/application_train.csv')\n",
    "test_control = pd.read_csv('../input/application_test.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "74c09643eee73b19d80acfca60dcd33df8747364"
   },
   "source": [
    "Fortunately, once we have taken the time to write a function, using it is simple (if there's a central theme in this notebook, it's use functions to make things simpler and reproducible!). The function above returns a `submission` dataframe we can upload to the competition, a `fi` dataframe of feature importances, and a `metrics` dataframe with validation and test performance. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {
    "_uuid": "3711f26b43863952e4edba7c3fc33f877ed499b2",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "submission, fi, metrics = model(train_control, test_control)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {
    "_uuid": "fed9ad5478719de194aeabec9b151f21effb99f5",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "metrics"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "4f3382baf72ccfc101660b2f062fee707e7cb041"
   },
   "source": [
    "The control slightly overfits because the training score is higher than the validation score. We can address this in later notebooks when we look at regularization (we already perform some regularization in this model by using `reg_lambda` and `reg_alpha` as well as early stopping). \n",
    "\n",
    "We can visualize the feature importance with another function, `plot_feature_importances`. The feature importances may be useful when it's time for feature selection. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {
    "_uuid": "c09f30bd4fb6e12b55d5a288fab358ffd1fe6f4d",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "fi_sorted = plot_feature_importances(fi)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {
    "_uuid": "6fe65cc25940a40c9966101acf9ea8f8477d7b77",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "submission.to_csv('control.csv', index = False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "e502ef191ea2201b05f85cece2202931744fa539"
   },
   "source": [
    "__The control scores 0.745 when submitted to the competition.__"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "392da17505a7feb3320242ff9785ac81346ee0b7"
   },
   "source": [
    "### Test One\n",
    "\n",
    "Let's conduct the first test. We will just need to pass in the data to the function, which does most of the work for us."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {
    "_uuid": "640bceff24f8c53aed68568c6ee1f70575adab69",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "submission_raw, fi_raw, metrics_raw = model(train, test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {
    "_uuid": "b404e436782d6de379bfa4756d58bed0175396f8",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "metrics_raw"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "0fcb737753ca178e32afda3fe29bf4c29262c379"
   },
   "source": [
    "Based on these numbers, the engineered features perform better than the control case. However, we will have to submit the predictions to the leaderboard before we can say if this better validation performance transfers to the testing data. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {
    "_uuid": "3313e261d947645ca0e72f824de45827a43de9c1",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "fi_raw_sorted = plot_feature_importances(fi_raw)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "3bf21f18e6430cea006cdd6b0139c52cb760d6ee"
   },
   "source": [
    "Examining the feature improtances, it looks as if a few of the feature we constructed are among the most important. Let's find the percentage of the top 100 most important features that we made in this notebook. However, rather than just compare to the original features, we need to compare to the _one-hot encoded_ original features. These are already recorded for us in `fi` (from the original data). "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {
    "_uuid": "b5abb37ffe89b767a01f494801c4db5bc3abc6db",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "top_100 = list(fi_raw_sorted['feature'])[:100]\n",
    "new_features = [x for x in top_100 if x not in list(fi['feature'])]\n",
    "\n",
    "print('%% of Top 100 Features created from the bureau data = %d.00' % len(new_features))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "4ae9e0771495bdfff89f221f435c5320c8fd27f0"
   },
   "source": [
    "Over half of the top 100 features were made by us! That should give us confidence that all the hard work we did was worthwhile. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {
    "_uuid": "0fe38b12afa98f44ad2a1e20da417f3e1f3728fd",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "submission_raw.to_csv('test_one.csv', index = False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "9bff36a22e3ee286f046f819bcca121c14df27a8"
   },
   "source": [
    "__Test one scores 0.759 when submitted to the competition.__"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "9469d920abeeb1c912391b06c1f4bd0175c15780"
   },
   "source": [
    "### Test Two\n",
    "\n",
    "That was easy, so let's do another run! Same as before but with the highly collinear variables removed. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {
    "_uuid": "008b1859c010277339ac2b1c2b52c0516c001774",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "submission_corrs, fi_corrs, metrics_corr = model(train_corrs_removed, test_corrs_removed)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {
    "_uuid": "6f91e580802f054053d497495de1c7aa4f4047eb",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "metrics_corr"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "c1c840d66cb195848d861133e24be78514624f30"
   },
   "source": [
    "These results are better than the control, but slightly lower than the raw features. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {
    "_uuid": "ed8a71b2d72719418dcb9463dea983dfe2b1ade8",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "fi_corrs_sorted = plot_feature_importances(fi_corrs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {
    "_uuid": "521e22fc1ba905c4851ca2d948260298a7527519",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "submission_corrs.to_csv('test_two.csv', index = False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "5cca331fe96700aa942ef56f96ce2fa34ad1933a"
   },
   "source": [
    "__Test Two scores 0.753 when submitted to the competition.__"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_uuid": "457467aab25e8c047a2c74a1392730b96203cfa2"
   },
   "source": [
    "# Results\n",
    "\n",
    "After all that work, we can say that including the extra information did improve performance! The model is definitely not optimized to our data, but we still had a noticeable improvement over the original dataset when using the calculated features. Let's officially summarize the performances:\n",
    "\n",
    "| __Experiment__ | __Train AUC__ | __Validation AUC__ | __Test AUC__  |\n",
    "|------------|-------|------------|-------|\n",
    "| __Control__    | 0.815 | 0.760      | 0.745 |\n",
    "| __Test One__   | 0.837 | 0.767      | 0.759 |\n",
    "| __Test Two__   | 0.826 | 0.765      | 0.753 |\n",
    "\n",
    "\n",
    "(Note that these scores may change from run to run of the notebook. I have not observed that the general ordering changes however.)\n",
    "\n",
    "All of our hard work translates to a small improvement of 0.014 ROC AUC over the original testing data. Removing the highly collinear variables slightly decreases performance so we will want to consider a different method for feature selection. Moreover, we can say that some of the features we built are among the most important as judged by the model. \n",
    "\n",
    "In a competition such as this, even an improvement of this size is enough to move us up 100s of spots on the leaderboard. By making numerous small improvements such as in this notebook, we can gradually achieve better and better performance. I encourage others to use the results here to make their own improvements, and I will continue to document the steps I take to help others. \n",
    "\n",
    "## Next Steps\n",
    "\n",
    "Going forward, we can now use the functions we developed in this notebook on the other datasets. There are still 4 other data files to use in our model! In the next notebook, we will incorporate the information from these other data files (which contain information on previous loans at Home Credit) into our training data. Then we can build the same model and run more experiments to determine the effect of our feature engineering. There is plenty more work to be done in this competition, and plenty more gains in performance to be had! I'll see you in the next notebook."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "_uuid": "94bdea7d6198e5cf61f073c3c4d0386d5961b564",
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}