ch11_金融和经济数据应用.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib\n",
    "matplotlib.rcParams['font.sans-serif'] = [u'SimHei']\n",
    "matplotlib.rcParams['axes.unicode_minus'] = False\n",
    "import os\n",
    "\n",
    "path =str(os.getcwd()).replace('\\\\','/')\n",
    "# s=pd.read_csv(path +'/datasets/ch08_data/tips.csv',engine = 'python')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "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>AAPL</th>\n",
       "      <th>JNJ</th>\n",
       "      <th>SPX</th>\n",
       "      <th>XOM</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2011-09-06</th>\n",
       "      <td>379.74</td>\n",
       "      <td>64.64</td>\n",
       "      <td>1165.24</td>\n",
       "      <td>71.15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2011-09-07</th>\n",
       "      <td>383.93</td>\n",
       "      <td>65.43</td>\n",
       "      <td>1198.62</td>\n",
       "      <td>73.65</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2011-09-08</th>\n",
       "      <td>384.14</td>\n",
       "      <td>64.95</td>\n",
       "      <td>1185.90</td>\n",
       "      <td>72.82</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2011-09-09</th>\n",
       "      <td>377.48</td>\n",
       "      <td>63.64</td>\n",
       "      <td>1154.23</td>\n",
       "      <td>71.01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2011-09-12</th>\n",
       "      <td>379.94</td>\n",
       "      <td>63.59</td>\n",
       "      <td>1162.27</td>\n",
       "      <td>71.84</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2011-09-13</th>\n",
       "      <td>384.62</td>\n",
       "      <td>63.61</td>\n",
       "      <td>1172.87</td>\n",
       "      <td>71.65</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2011-09-14</th>\n",
       "      <td>389.30</td>\n",
       "      <td>63.73</td>\n",
       "      <td>1188.68</td>\n",
       "      <td>72.64</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              AAPL    JNJ      SPX    XOM\n",
       "2011-09-06  379.74  64.64  1165.24  71.15\n",
       "2011-09-07  383.93  65.43  1198.62  73.65\n",
       "2011-09-08  384.14  64.95  1185.90  72.82\n",
       "2011-09-09  377.48  63.64  1154.23  71.01\n",
       "2011-09-12  379.94  63.59  1162.27  71.84\n",
       "2011-09-13  384.62  63.61  1172.87  71.65\n",
       "2011-09-14  389.30  63.73  1188.68  72.64"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "prices=pd.read_csv(path +'/datasets/ch11_data/stock_px.csv',engine = 'python',parse_dates=True,index_col=0)\n",
    "prices=prices.loc['2011-09-06':'2011-09-14',['AAPL','JNJ','SPX','XOM']]\n",
    "prices"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "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>AAPL</th>\n",
       "      <th>JNJ</th>\n",
       "      <th>XOM</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2011-09-06</th>\n",
       "      <td>18173500.0</td>\n",
       "      <td>15848300.0</td>\n",
       "      <td>25416300.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2011-09-07</th>\n",
       "      <td>12492000.0</td>\n",
       "      <td>10759700.0</td>\n",
       "      <td>23108400.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2011-09-08</th>\n",
       "      <td>14839800.0</td>\n",
       "      <td>15551500.0</td>\n",
       "      <td>22434800.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2011-09-09</th>\n",
       "      <td>20171900.0</td>\n",
       "      <td>17008200.0</td>\n",
       "      <td>27969100.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2011-09-12</th>\n",
       "      <td>16697300.0</td>\n",
       "      <td>13448200.0</td>\n",
       "      <td>26205800.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  AAPL         JNJ         XOM\n",
       "2011-09-06  18173500.0  15848300.0  25416300.0\n",
       "2011-09-07  12492000.0  10759700.0  23108400.0\n",
       "2011-09-08  14839800.0  15551500.0  22434800.0\n",
       "2011-09-09  20171900.0  17008200.0  27969100.0\n",
       "2011-09-12  16697300.0  13448200.0  26205800.0"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "volume=pd.read_csv(path +'/datasets/ch11_data/volume.csv',engine = 'python',parse_dates=True,index_col=0)\n",
    "volume=volume.loc['2011-09-06':'2011-09-12',['AAPL','JNJ','XOM']]\n",
    "volume"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "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>AAPL</th>\n",
       "      <th>JNJ</th>\n",
       "      <th>SPX</th>\n",
       "      <th>XOM</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2011-09-06</th>\n",
       "      <td>6.901205e+09</td>\n",
       "      <td>1.024434e+09</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.808370e+09</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2011-09-07</th>\n",
       "      <td>4.796054e+09</td>\n",
       "      <td>7.040072e+08</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.701934e+09</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2011-09-08</th>\n",
       "      <td>5.700561e+09</td>\n",
       "      <td>1.010070e+09</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.633702e+09</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2011-09-09</th>\n",
       "      <td>7.614489e+09</td>\n",
       "      <td>1.082402e+09</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.986086e+09</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2011-09-12</th>\n",
       "      <td>6.343972e+09</td>\n",
       "      <td>8.551710e+08</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.882625e+09</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2011-09-13</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2011-09-14</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                    AAPL           JNJ  SPX           XOM\n",
       "2011-09-06  6.901205e+09  1.024434e+09  NaN  1.808370e+09\n",
       "2011-09-07  4.796054e+09  7.040072e+08  NaN  1.701934e+09\n",
       "2011-09-08  5.700561e+09  1.010070e+09  NaN  1.633702e+09\n",
       "2011-09-09  7.614489e+09  1.082402e+09  NaN  1.986086e+09\n",
       "2011-09-12  6.343972e+09  8.551710e+08  NaN  1.882625e+09\n",
       "2011-09-13           NaN           NaN  NaN           NaN\n",
       "2011-09-14           NaN           NaN  NaN           NaN"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 计算一个成交量加权平均价格(为了简单起见，假设成交量数据是价值数据的子集)\n",
    "prices*volume"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "AAPL    380.655181\n",
       "JNJ      64.394769\n",
       "SPX            NaN\n",
       "XOM      72.024288\n",
       "dtype: float64"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "vwap = (prices*volume).sum()/volume.sum()\n",
    "vwap"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "AAPL    380.655181\n",
       "JNJ      64.394769\n",
       "XOM      72.024288\n",
       "dtype: float64"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "vwap.dropna()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(              AAPL    JNJ    XOM\n",
       " 2011-09-06  379.74  64.64  71.15\n",
       " 2011-09-07  383.93  65.43  73.65\n",
       " 2011-09-08  384.14  64.95  72.82\n",
       " 2011-09-09  377.48  63.64  71.01\n",
       " 2011-09-12  379.94  63.59  71.84,\n",
       "                   AAPL         JNJ         XOM\n",
       " 2011-09-06  18173500.0  15848300.0  25416300.0\n",
       " 2011-09-07  12492000.0  10759700.0  23108400.0\n",
       " 2011-09-08  14839800.0  15551500.0  22434800.0\n",
       " 2011-09-09  20171900.0  17008200.0  27969100.0\n",
       " 2011-09-12  16697300.0  13448200.0  26205800.0)"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# SPX在volume中找不到，可以直接丢弃。\n",
    "# 如果希望手工对齐，可以使用DataFrame的align方法，它返回的是一个元组，含有两个对象的重索引版本\n",
    "prices.align(volume,join='inner')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "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>one</th>\n",
       "      <th>three</th>\n",
       "      <th>two</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>s2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b</th>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>s2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>c</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>s2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>f</th>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>s2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   one  three two\n",
       "a  0.0    1.0  s2\n",
       "b  1.0    NaN  s2\n",
       "c  2.0    2.0  s2\n",
       "f  NaN    0.0  s2"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 通过一组可能不同的索引的Series构建一个DataFrame\n",
    "s1 = pd.Series(range(3),index=['a','b','c'])\n",
    "s2 = pd.Series(range(4),index=['d','b','c','e'])\n",
    "s3 = pd.Series(range(3),index=['f','a','c'])\n",
    "pd.DataFrame({'one':s1,'two':'s2','three':s3})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "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>one</th>\n",
       "      <th>three</th>\n",
       "      <th>two</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>f</th>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>c</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>e</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   one  three  two\n",
       "f  NaN    0.0  NaN\n",
       "a  0.0    1.0  NaN\n",
       "c  2.0    2.0  2.0\n",
       "e  NaN    NaN  3.0"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 也可以显示定义结果的索引（丢弃其余数据）\n",
    "pd.DataFrame({'one':s1,'two':s2,'three':s3},index=list('face'))  # 只保留部分索引的数据"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 频率不同的时间序列的运算\n",
    "频率转换和重对齐的两大工具是resample和reindex\n",
    "- resample用于将数据转换到固定频率\n",
    "- reindex用于使数据符合一个新索引\n",
    "- 都支持插值（如前向填充）逻辑"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2012-06-13    1.063746\n",
       "2012-06-20   -0.417105\n",
       "2012-06-27    1.050647\n",
       "Freq: W-WED, dtype: float64"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 周型时间序列\n",
    "ts1 = pd.Series(np.random.randn(3),index=pd.date_range('2012-6-13',periods=3,freq='W-WED'))\n",
    "ts1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2012-06-13    1.063746\n",
       "2012-06-14         NaN\n",
       "2012-06-15         NaN\n",
       "2012-06-18         NaN\n",
       "2012-06-19         NaN\n",
       "2012-06-20   -0.417105\n",
       "2012-06-21         NaN\n",
       "2012-06-22         NaN\n",
       "2012-06-25         NaN\n",
       "2012-06-26         NaN\n",
       "2012-06-27    1.050647\n",
       "Freq: B, dtype: float64"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 如果将其重采样到工作日（星期一到星期五）频率，则那些没有数据的日子就会出现一个“空洞”\n",
    "ts1.resample('B').asfreq()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "D:\\Python\\anaconda\\lib\\site-packages\\ipykernel_launcher.py:2: FutureWarning: fill_method is deprecated to .resample()\n",
      "the new syntax is .resample(...).ffill()\n",
      "  \n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "2012-06-13    1.063746\n",
       "2012-06-14    1.063746\n",
       "2012-06-15    1.063746\n",
       "2012-06-18    1.063746\n",
       "2012-06-19    1.063746\n",
       "2012-06-20   -0.417105\n",
       "2012-06-21   -0.417105\n",
       "2012-06-22   -0.417105\n",
       "2012-06-25   -0.417105\n",
       "2012-06-26   -0.417105\n",
       "2012-06-27    1.050647\n",
       "Freq: B, dtype: float64"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 只需将fill_method设置为‘ffill’即可用起那么的值填充这些空白\n",
    "ts1.resample('B',fill_method='ffill')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2012-06-12   -0.770483\n",
       "2012-06-17    0.319822\n",
       "2012-06-18   -1.141505\n",
       "2012-06-21    1.728708\n",
       "2012-06-22    1.991976\n",
       "2012-06-29    2.072726\n",
       "dtype: float64"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 将较低频率的数据升采样到较高的规整频率是一种不错的解决方案\n",
    "# 但对于不规整时间序列，如下\n",
    "dates = pd.DatetimeIndex(['2012-6-12','2012-6-17','2012-6-18','2012-6-21','2012-6-22','2012-6-29'])\n",
    "ts2 = pd.Series(np.random.randn(6),index=dates)\n",
    "ts2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2012-06-12         NaN\n",
       "2012-06-17    1.063746\n",
       "2012-06-18    1.063746\n",
       "2012-06-21   -0.417105\n",
       "2012-06-22   -0.417105\n",
       "2012-06-29    1.050647\n",
       "dtype: float64"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 如果要将ts1中‘当前值’(即前向填充)加到ts2.\n",
    "# 一个办法是将两者重采用为规整频率后再相加，但是，如果想维持ts2的日期索引，则reindex是一种更好的解决方法\n",
    "ts1.reindex(ts2.index,method='ffill')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2012-06-12         NaN\n",
       "2012-06-17    1.383567\n",
       "2012-06-18   -0.077759\n",
       "2012-06-21    1.311603\n",
       "2012-06-22    1.574871\n",
       "2012-06-29    3.123373\n",
       "dtype: float64"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ts2+ts1.reindex(ts2.index,method='ffill')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 使用Period\n",
    "Period（表示时间区间）提供另一种处理不同频率时间序列的方法，尤其是那些有着特殊规范的以年或季度为频率的金融或经济序列。比如：一个公司可能会发布其以6月结尾的财年的每季度盈利报告，即频率为Q-JUN。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1984Q2    1.78\n",
       "1984Q3    1.94\n",
       "1984Q4    2.08\n",
       "1985Q1    2.01\n",
       "1985Q2    2.15\n",
       "1985Q3    2.31\n",
       "1985Q4    2.46\n",
       "Freq: Q-SEP, dtype: float64"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 以下是两个有关GDP和通货膨胀的宏观经济时间序列\n",
    "gdp = pd.Series([1.78,1.94,2.08,2.01,2.15,2.31,2.46],\n",
    "                index = pd.period_range('1984Q2',periods=7,freq='Q-SEP'))\n",
    "infl = pd.Series([0.025,0.045,0.037,0.04],\n",
    "                 index = pd.period_range('1982',periods=4,freq='A-DEC'))\n",
    "gdp"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1982    0.025\n",
       "1983    0.045\n",
       "1984    0.037\n",
       "1985    0.040\n",
       "Freq: A-DEC, dtype: float64"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "infl"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1983Q1    0.025\n",
       "1984Q1    0.045\n",
       "1985Q1    0.037\n",
       "1986Q1    0.040\n",
       "Freq: Q-SEP, dtype: float64"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 跟Timestamp的时间序列不同，由Period索引的两个不同频率的时间序列之间的运算必须进行显式转换\n",
    "# infl值是每年年末观测的，可以将其转换到Q-SEP以得到该频率下的正确时期\n",
    "infl_q = infl.asfreq('Q-SEP',how='end')\n",
    "infl_q"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1984Q2    0.045\n",
       "1984Q3    0.045\n",
       "1984Q4    0.045\n",
       "1985Q1    0.037\n",
       "1985Q2    0.037\n",
       "1985Q3    0.037\n",
       "1985Q4    0.037\n",
       "Freq: Q-SEP, dtype: float64"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 然后这个时间序列就可以被重索引了（使用前向填充以匹配gdp）\n",
    "infl_q.reindex(gdp.index,method='ffill')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 时间和“最当前”数据选取\n",
    "从数据中抽取每天特定时间的数据"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DatetimeIndex(['2012-06-01 09:30:00', '2012-06-01 09:31:00',\n",
       "               '2012-06-01 09:32:00', '2012-06-01 09:33:00',\n",
       "               '2012-06-01 09:34:00', '2012-06-01 09:35:00',\n",
       "               '2012-06-01 09:36:00', '2012-06-01 09:37:00',\n",
       "               '2012-06-01 09:38:00', '2012-06-01 09:39:00',\n",
       "               ...\n",
       "               '2012-06-01 15:50:00', '2012-06-01 15:51:00',\n",
       "               '2012-06-01 15:52:00', '2012-06-01 15:53:00',\n",
       "               '2012-06-01 15:54:00', '2012-06-01 15:55:00',\n",
       "               '2012-06-01 15:56:00', '2012-06-01 15:57:00',\n",
       "               '2012-06-01 15:58:00', '2012-06-01 15:59:00'],\n",
       "              dtype='datetime64[ns]', length=390, freq='T')"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 生成一个交易日内的日期范围和时间序列\n",
    "rng = pd.date_range('2012-06-01 09:30','2012-06-01 15:59',freq='T')  #freq='T':按分钟\n",
    "rng"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DatetimeIndex(['2012-06-01 09:30:00', '2012-06-01 09:31:00',\n",
       "               '2012-06-01 09:32:00', '2012-06-01 09:33:00',\n",
       "               '2012-06-01 09:34:00', '2012-06-01 09:35:00',\n",
       "               '2012-06-01 09:36:00', '2012-06-01 09:37:00',\n",
       "               '2012-06-01 09:38:00', '2012-06-01 09:39:00',\n",
       "               ...\n",
       "               '2012-06-06 15:50:00', '2012-06-06 15:51:00',\n",
       "               '2012-06-06 15:52:00', '2012-06-06 15:53:00',\n",
       "               '2012-06-06 15:54:00', '2012-06-06 15:55:00',\n",
       "               '2012-06-06 15:56:00', '2012-06-06 15:57:00',\n",
       "               '2012-06-06 15:58:00', '2012-06-06 15:59:00'],\n",
       "              dtype='datetime64[ns]', length=1560, freq=None)"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 生成4天的时间点（9:30 ~ 15:59 之间的值）\n",
    "# pd.offsets.BDay(1)：添加一个工作日\n",
    "rng = rng.append([rng + pd.offsets.BDay(i) for i in range(1,4)])  # 增加三个工作日：6.4,6.5,6.6\n",
    "rng"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2012-06-01 09:30:00       0.0\n",
       "2012-06-01 09:31:00       1.0\n",
       "2012-06-01 09:32:00       2.0\n",
       "2012-06-01 09:33:00       3.0\n",
       "2012-06-01 09:34:00       4.0\n",
       "2012-06-01 09:35:00       5.0\n",
       "2012-06-01 09:36:00       6.0\n",
       "2012-06-01 09:37:00       7.0\n",
       "2012-06-01 09:38:00       8.0\n",
       "2012-06-01 09:39:00       9.0\n",
       "2012-06-01 09:40:00      10.0\n",
       "2012-06-01 09:41:00      11.0\n",
       "2012-06-01 09:42:00      12.0\n",
       "2012-06-01 09:43:00      13.0\n",
       "2012-06-01 09:44:00      14.0\n",
       "2012-06-01 09:45:00      15.0\n",
       "2012-06-01 09:46:00      16.0\n",
       "2012-06-01 09:47:00      17.0\n",
       "2012-06-01 09:48:00      18.0\n",
       "2012-06-01 09:49:00      19.0\n",
       "2012-06-01 09:50:00      20.0\n",
       "2012-06-01 09:51:00      21.0\n",
       "2012-06-01 09:52:00      22.0\n",
       "2012-06-01 09:53:00      23.0\n",
       "2012-06-01 09:54:00      24.0\n",
       "2012-06-01 09:55:00      25.0\n",
       "2012-06-01 09:56:00      26.0\n",
       "2012-06-01 09:57:00      27.0\n",
       "2012-06-01 09:58:00      28.0\n",
       "2012-06-01 09:59:00      29.0\n",
       "                        ...  \n",
       "2012-06-06 15:30:00    1530.0\n",
       "2012-06-06 15:31:00    1531.0\n",
       "2012-06-06 15:32:00    1532.0\n",
       "2012-06-06 15:33:00    1533.0\n",
       "2012-06-06 15:34:00    1534.0\n",
       "2012-06-06 15:35:00    1535.0\n",
       "2012-06-06 15:36:00    1536.0\n",
       "2012-06-06 15:37:00    1537.0\n",
       "2012-06-06 15:38:00    1538.0\n",
       "2012-06-06 15:39:00    1539.0\n",
       "2012-06-06 15:40:00    1540.0\n",
       "2012-06-06 15:41:00    1541.0\n",
       "2012-06-06 15:42:00    1542.0\n",
       "2012-06-06 15:43:00    1543.0\n",
       "2012-06-06 15:44:00    1544.0\n",
       "2012-06-06 15:45:00    1545.0\n",
       "2012-06-06 15:46:00    1546.0\n",
       "2012-06-06 15:47:00    1547.0\n",
       "2012-06-06 15:48:00    1548.0\n",
       "2012-06-06 15:49:00    1549.0\n",
       "2012-06-06 15:50:00    1550.0\n",
       "2012-06-06 15:51:00    1551.0\n",
       "2012-06-06 15:52:00    1552.0\n",
       "2012-06-06 15:53:00    1553.0\n",
       "2012-06-06 15:54:00    1554.0\n",
       "2012-06-06 15:55:00    1555.0\n",
       "2012-06-06 15:56:00    1556.0\n",
       "2012-06-06 15:57:00    1557.0\n",
       "2012-06-06 15:58:00    1558.0\n",
       "2012-06-06 15:59:00    1559.0\n",
       "Length: 1560, dtype: float64"
      ]
     },
     "execution_count": 76,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ts = pd.Series(np.arange(len(rng),dtype=float),index=rng)\n",
    "ts"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2012-06-01 10:00:00      30.0\n",
       "2012-06-04 10:00:00     420.0\n",
       "2012-06-05 10:00:00     810.0\n",
       "2012-06-06 10:00:00    1200.0\n",
       "dtype: float64"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 利用python的datetime.time对象进行索引即可抽取出这些点上的值\n",
    "from datetime import time\n",
    "ts[time(10,0)]  #  抽取10点的数据"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2012-06-01 10:00:00      30.0\n",
       "2012-06-04 10:00:00     420.0\n",
       "2012-06-05 10:00:00     810.0\n",
       "2012-06-06 10:00:00    1200.0\n",
       "dtype: float64"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 实际上，该操作用到了实例方法at_time（各时间序列以及类似的DataFrame对象都有）\n",
    "ts.at_time(time(10,0))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2012-06-01 10:00:00      30.0\n",
       "2012-06-01 10:01:00      31.0\n",
       "2012-06-04 10:00:00     420.0\n",
       "2012-06-04 10:01:00     421.0\n",
       "2012-06-05 10:00:00     810.0\n",
       "2012-06-05 10:01:00     811.0\n",
       "2012-06-06 10:00:00    1200.0\n",
       "2012-06-06 10:01:00    1201.0\n",
       "dtype: float64"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# between_time方法是用于选取两个Time对象之间的值\n",
    "ts.between_time(time(10,0),time(10,1))  # 抽取出10点到10点零一分钟之间的值，包含两个边界的数据"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([   0,    1,    2,    4,    5,    7,    8,   10,   12,   13,   15,\n",
       "         16,   17,   19,   20,   21,   26,   27,   28,   29,   31,   32,\n",
       "         33,   35,   36,   38,   44,   47,   48,   49,   50,   54,   55,\n",
       "         63,   67,   68,   69,   73,   74,   76,   78,   80,   82,   84,\n",
       "         86,   89,   91,   93,   95,   98,  101,  109,  114,  122,  124,\n",
       "        125,  129,  131,  132,  133,  137,  139,  141,  142,  143,  145,\n",
       "        147,  152,  153,  154,  155,  156,  159,  160,  162,  165,  167,\n",
       "        168,  169,  170,  171,  172,  173,  176,  177,  178,  182,  183,\n",
       "        184,  185,  186,  187,  189,  190,  192,  193,  197,  202,  203,\n",
       "        204,  205,  206,  207,  209,  210,  211,  212,  213,  215,  216,\n",
       "        219,  221,  223,  224,  225,  226,  229,  231,  232,  236,  238,\n",
       "        242,  245,  247,  249,  250,  251,  252,  253,  254,  255,  257,\n",
       "        258,  260,  261,  263,  264,  265,  268,  272,  273,  275,  277,\n",
       "        281,  282,  284,  286,  287,  288,  292,  294,  297,  305,  306,\n",
       "        307,  308,  311,  313,  314,  315,  317,  318,  319,  320,  321,\n",
       "        322,  323,  329,  331,  332,  334,  339,  341,  344,  346,  347,\n",
       "        349,  351,  353,  355,  356,  361,  362,  363,  367,  368,  370,\n",
       "        372,  374,  376,  379,  380,  381,  382,  383,  384,  389,  390,\n",
       "        391,  392,  394,  397,  398,  399,  401,  404,  406,  407,  408,\n",
       "        409,  410,  411,  412,  415,  419,  421,  423,  426,  427,  429,\n",
       "        431,  432,  433,  434,  435,  436,  438,  439,  440,  441,  443,\n",
       "        446,  447,  449,  450,  451,  454,  456,  457,  458,  459,  461,\n",
       "        463,  465,  467,  469,  470,  471,  474,  476,  477,  480,  481,\n",
       "        482,  485,  486,  489,  490,  492,  493,  494,  496,  498,  499,\n",
       "        500,  501,  502,  505,  506,  507,  510,  511,  512,  515,  516,\n",
       "        518,  520,  521,  523,  525,  529,  531,  533,  534,  535,  536,\n",
       "        540,  541,  543,  544,  545,  546,  549,  550,  551,  552,  555,\n",
       "        561,  562,  564,  565,  566,  567,  568,  569,  570,  574,  575,\n",
       "        577,  578,  579,  581,  584,  586,  588,  591,  593,  598,  599,\n",
       "        601,  605,  606,  608,  609,  610,  611,  612,  613,  615,  618,\n",
       "        620,  623,  624,  626,  630,  631,  634,  635,  636,  637,  638,\n",
       "        639,  640,  641,  644,  646,  648,  650,  651,  652,  654,  655,\n",
       "        658,  660,  661,  663,  666,  669,  671,  672,  673,  674,  675,\n",
       "        676,  678,  681,  682,  683,  685,  690,  691,  692,  693,  697,\n",
       "        698,  699,  700,  701,  703,  704,  706,  707,  709,  710,  711,\n",
       "        712,  713,  717,  718,  719,  721,  722,  723,  725,  727,  728,\n",
       "        730,  732,  733,  736,  737,  741,  742,  743,  744,  745,  751,\n",
       "        752,  753,  754,  755,  756,  757,  758,  759,  760,  761,  762,\n",
       "        767,  768,  771,  772,  775,  777,  778,  780,  781,  791,  792,\n",
       "        793,  794,  797,  799,  801,  803,  804,  807,  809,  811,  813,\n",
       "        816,  818,  819,  821,  823,  824,  826,  827,  829,  830,  831,\n",
       "        832,  834,  835,  838,  840,  843,  847,  850,  852,  854,  860,\n",
       "        863,  864,  865,  867,  868,  869,  870,  871,  872,  874,  876,\n",
       "        877,  884,  885,  887,  889,  890,  891,  892,  895,  896,  897,\n",
       "        899,  900,  902,  903,  904,  905,  910,  911,  912,  913,  915,\n",
       "        918,  919,  924,  930,  933,  937,  938,  940,  941,  943,  945,\n",
       "        948,  949,  953,  954,  955,  956,  957,  961,  962,  963,  964,\n",
       "        965,  966,  968,  969,  970,  972,  974,  975,  976,  977,  978,\n",
       "        979,  980,  984,  987,  988,  989,  991,  992,  993,  995,  996,\n",
       "       1000, 1001, 1003, 1004, 1005, 1008, 1011, 1014, 1016, 1019, 1020,\n",
       "       1021, 1022, 1024, 1026, 1031, 1033, 1034, 1038, 1041, 1043, 1045,\n",
       "       1049, 1050, 1051, 1052, 1055, 1056, 1057, 1058, 1060, 1061, 1063,\n",
       "       1064, 1066, 1068, 1071, 1072, 1073, 1075, 1076, 1078, 1079, 1080,\n",
       "       1081, 1083, 1084, 1088, 1089, 1092, 1095, 1098, 1099, 1101, 1103,\n",
       "       1104, 1106, 1107, 1109, 1110, 1112, 1115, 1116, 1117, 1119, 1122,\n",
       "       1124, 1128, 1129, 1135, 1137, 1138, 1139, 1142, 1143, 1144, 1145,\n",
       "       1150, 1151, 1154, 1156, 1157, 1159, 1160, 1163, 1164, 1166, 1167,\n",
       "       1168, 1169, 1170, 1171, 1173, 1174, 1175, 1177, 1179, 1180, 1182,\n",
       "       1183, 1184, 1186, 1187, 1190, 1191, 1192, 1194, 1195, 1196, 1198,\n",
       "       1201, 1202, 1203, 1204, 1207, 1208, 1209, 1210, 1212, 1213, 1215,\n",
       "       1217, 1218, 1219, 1222, 1223, 1224, 1225, 1226, 1227, 1228, 1229,\n",
       "       1231, 1233, 1234, 1235, 1236, 1238, 1239, 1243, 1246, 1248, 1252,\n",
       "       1257, 1259, 1261, 1262, 1263, 1265, 1266, 1267, 1268, 1270, 1272,\n",
       "       1273, 1274, 1275, 1278, 1283, 1284, 1287, 1292, 1293, 1295, 1296,\n",
       "       1297, 1302, 1303, 1304, 1305, 1308, 1309, 1310, 1311, 1312, 1313,\n",
       "       1314, 1315, 1317, 1318, 1321, 1322, 1323, 1325, 1326, 1327, 1328,\n",
       "       1332, 1335, 1337, 1342, 1345, 1346, 1348, 1349, 1350, 1351, 1352,\n",
       "       1353, 1354, 1359, 1361, 1362, 1363, 1364, 1365, 1366, 1368, 1370,\n",
       "       1371, 1373, 1377, 1378, 1379, 1380, 1381, 1383, 1384, 1385, 1386,\n",
       "       1387, 1388, 1389, 1390, 1393, 1394, 1395, 1397, 1399, 1400, 1401,\n",
       "       1403, 1404, 1405, 1406, 1407, 1408, 1409, 1413, 1416, 1417, 1419,\n",
       "       1421, 1422, 1425, 1427, 1429, 1430, 1432, 1435, 1438, 1439, 1441,\n",
       "       1444, 1445, 1447, 1448, 1450, 1452, 1454, 1455, 1456, 1458, 1463,\n",
       "       1465, 1468, 1470, 1472, 1473, 1475, 1477, 1481, 1482, 1483, 1485,\n",
       "       1486, 1487, 1492, 1497, 1498, 1499, 1500, 1502, 1503, 1506, 1507,\n",
       "       1511, 1512, 1513, 1514, 1517, 1518, 1521, 1522, 1524, 1526, 1527,\n",
       "       1529, 1530, 1531, 1532, 1533, 1534, 1536, 1537, 1540, 1541, 1542,\n",
       "       1543, 1544, 1545, 1547, 1549, 1550, 1551, 1552, 1553, 1554, 1556,\n",
       "       1557, 1559])"
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 可能刚好就没有任何数据落在某个具体时间上（比如上午10点），这时，希望得到上午10点之前最后出现的那个值\n",
    "# 将该时间序列的大部分内容随机设置为NA\n",
    "indexer = np.sort(np.random.permutation(len(ts))[700:])\n",
    "indexer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2012-06-01 09:50:00     NaN\n",
       "2012-06-01 09:51:00     NaN\n",
       "2012-06-01 09:52:00    22.0\n",
       "2012-06-01 09:53:00    23.0\n",
       "2012-06-01 09:54:00    24.0\n",
       "2012-06-01 09:55:00    25.0\n",
       "2012-06-01 09:56:00     NaN\n",
       "2012-06-01 09:57:00     NaN\n",
       "2012-06-01 09:58:00     NaN\n",
       "2012-06-01 09:59:00     NaN\n",
       "2012-06-01 10:00:00    30.0\n",
       "dtype: float64"
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "irr_ts = ts.copy()\n",
    "irr_ts[indexer]=np.nan\n",
    "irr_ts['2012-06-01 09:50':'2012-06-01 10:00']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DatetimeIndex(['2012-06-01 10:00:00', '2012-06-04 10:00:00',\n",
       "               '2012-06-05 10:00:00', '2012-06-06 10:00:00'],\n",
       "              dtype='datetime64[ns]', freq='B')"
      ]
     },
     "execution_count": 84,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 如果将一组Timestamp传入asof方法，就能得到这些时间点处（或其之前最近）的有效值（非NA）\n",
    "# 构造一个日期范围（每天上午10点），然后将其传入asof\n",
    "selection = pd.date_range('2012-06-01 10:00',periods=4,freq='B')\n",
    "selection"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2012-06-01 10:00:00      30.0\n",
       "2012-06-04 10:00:00     420.0\n",
       "2012-06-05 10:00:00     810.0\n",
       "2012-06-06 10:00:00    1200.0\n",
       "Freq: B, dtype: float64"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "irr_ts.asof(selection)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 拼接多个数据源\n",
    "- 在一个特定的时间点上，从一个数据源切换到另一个数据源\n",
    "- 用另一个时间序列对当前时间序列中的缺失值“打补丁”\n",
    "- 将数据中的符号（国家、资产代码等）替换为实际数据"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "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>a</th>\n",
       "      <th>b</th>\n",
       "      <th>c</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2012-06-12</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-13</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-14</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-15</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-16</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-17</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              a    b    c\n",
       "2012-06-12  1.0  1.0  1.0\n",
       "2012-06-13  1.0  1.0  1.0\n",
       "2012-06-14  1.0  1.0  1.0\n",
       "2012-06-15  1.0  1.0  1.0\n",
       "2012-06-16  1.0  1.0  1.0\n",
       "2012-06-17  1.0  1.0  1.0"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 在特定时刻从一个时间序列切换到另一个，用concat将两个TimeSeries或DataFrame对象合并到一起\n",
    "data1 = pd.DataFrame(np.ones((6,3),dtype=float),\n",
    "                     columns=['a','b','c'],\n",
    "                     index = pd.date_range('6/12/2012',periods=6))\n",
    "data1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "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>a</th>\n",
       "      <th>b</th>\n",
       "      <th>c</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2012-06-13</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-14</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-15</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-16</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-17</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-18</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              a    b    c\n",
       "2012-06-13  2.0  2.0  2.0\n",
       "2012-06-14  2.0  2.0  2.0\n",
       "2012-06-15  2.0  2.0  2.0\n",
       "2012-06-16  2.0  2.0  2.0\n",
       "2012-06-17  2.0  2.0  2.0\n",
       "2012-06-18  2.0  2.0  2.0"
      ]
     },
     "execution_count": 91,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data2 = pd.DataFrame(np.ones((6,3),dtype=float)*2,\n",
    "                     columns=['a','b','c'],\n",
    "                     index = pd.date_range('6/13/2012',periods=6))\n",
    "data2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "D:\\Python\\anaconda\\lib\\site-packages\\ipykernel_launcher.py:1: DeprecationWarning: \n",
      ".ix is deprecated. Please use\n",
      ".loc for label based indexing or\n",
      ".iloc for positional indexing\n",
      "\n",
      "See the documentation here:\n",
      "http://pandas.pydata.org/pandas-docs/stable/indexing.html#ix-indexer-is-deprecated\n",
      "  \"\"\"Entry point for launching an IPython kernel.\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>a</th>\n",
       "      <th>b</th>\n",
       "      <th>c</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2012-06-12</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-13</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-14</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-15</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-16</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-17</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-18</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              a    b    c\n",
       "2012-06-12  1.0  1.0  1.0\n",
       "2012-06-13  1.0  1.0  1.0\n",
       "2012-06-14  1.0  1.0  1.0\n",
       "2012-06-15  2.0  2.0  2.0\n",
       "2012-06-16  2.0  2.0  2.0\n",
       "2012-06-17  2.0  2.0  2.0\n",
       "2012-06-18  2.0  2.0  2.0"
      ]
     },
     "execution_count": 92,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "spliced = pd.concat([data1.ix[:'2012-06-14'],data2.ix['2012-06-15':]])\n",
    "spliced"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "D:\\Python\\anaconda\\lib\\site-packages\\ipykernel_launcher.py:5: DeprecationWarning: \n",
      ".ix is deprecated. Please use\n",
      ".loc for label based indexing or\n",
      ".iloc for positional indexing\n",
      "\n",
      "See the documentation here:\n",
      "http://pandas.pydata.org/pandas-docs/stable/indexing.html#ix-indexer-is-deprecated\n",
      "  \"\"\"\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>a</th>\n",
       "      <th>b</th>\n",
       "      <th>c</th>\n",
       "      <th>d</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2012-06-12</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-13</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-14</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-15</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-16</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-17</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-18</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              a    b    c    d\n",
       "2012-06-12  1.0  1.0  1.0  NaN\n",
       "2012-06-13  1.0  1.0  1.0  NaN\n",
       "2012-06-14  1.0  1.0  1.0  NaN\n",
       "2012-06-15  2.0  2.0  2.0  2.0\n",
       "2012-06-16  2.0  2.0  2.0  2.0\n",
       "2012-06-17  2.0  2.0  2.0  2.0\n",
       "2012-06-18  2.0  2.0  2.0  2.0"
      ]
     },
     "execution_count": 94,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 假设data1缺失了data2中存在的某个时间序列\n",
    "data2 = pd.DataFrame(np.ones((6,4),dtype=float)*2,\n",
    "                     columns=['a','b','c','d'],\n",
    "                     index = pd.date_range('6/13/2012',periods=6))\n",
    "spliced = pd.concat([data1.ix[:'2012-06-14'],data2.ix['2012-06-15':]])\n",
    "spliced"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "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>a</th>\n",
       "      <th>b</th>\n",
       "      <th>c</th>\n",
       "      <th>d</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2012-06-12</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-13</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-14</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-15</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-16</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-17</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-18</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              a    b    c    d\n",
       "2012-06-12  1.0  1.0  1.0  NaN\n",
       "2012-06-13  1.0  1.0  1.0  2.0\n",
       "2012-06-14  1.0  1.0  1.0  2.0\n",
       "2012-06-15  2.0  2.0  2.0  2.0\n",
       "2012-06-16  2.0  2.0  2.0  2.0\n",
       "2012-06-17  2.0  2.0  2.0  2.0\n",
       "2012-06-18  2.0  2.0  2.0  2.0"
      ]
     },
     "execution_count": 95,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# combine_first可以引入合并点之前的数据，这样也就扩展了‘d’项的历史\n",
    "spliced_filled = spliced.combine_first(data2)\n",
    "spliced_filled"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "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>a</th>\n",
       "      <th>b</th>\n",
       "      <th>c</th>\n",
       "      <th>d</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2012-06-12</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-13</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-14</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-15</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-16</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-17</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-18</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              a    b    c    d\n",
       "2012-06-12  1.0  1.0  1.0  NaN\n",
       "2012-06-13  1.0  1.0  1.0  2.0\n",
       "2012-06-14  1.0  1.0  1.0  2.0\n",
       "2012-06-15  2.0  2.0  2.0  2.0\n",
       "2012-06-16  2.0  2.0  2.0  2.0\n",
       "2012-06-17  2.0  2.0  2.0  2.0\n",
       "2012-06-18  2.0  2.0  2.0  2.0"
      ]
     },
     "execution_count": 97,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# DataFrame也有一个类似方法update，可以实现就地更新，如果只想填充空洞，则必须传入overwrite=False才行\n",
    "spliced.update(data2,overwrite=False)\n",
    "spliced"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "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>a</th>\n",
       "      <th>b</th>\n",
       "      <th>c</th>\n",
       "      <th>d</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2012-06-12</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-13</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-14</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-15</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-16</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-17</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-18</th>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              a    b    c    d\n",
       "2012-06-12  1.0  1.0  1.0  NaN\n",
       "2012-06-13  1.0  1.0  1.0  2.0\n",
       "2012-06-14  1.0  1.0  1.0  2.0\n",
       "2012-06-15  1.0  2.0  1.0  2.0\n",
       "2012-06-16  1.0  2.0  1.0  2.0\n",
       "2012-06-17  1.0  2.0  1.0  2.0\n",
       "2012-06-18  NaN  2.0  NaN  2.0"
      ]
     },
     "execution_count": 98,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 上面的都可以实现将数据中的符号替换为实际数据，但有时利用DataFrame的索引机制直接对列进行设置会更简单\n",
    "cp_spliced = spliced.copy()\n",
    "cp_spliced[['a','c']] = data1[['a','c']]\n",
    "cp_spliced"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 收益指数和累计指数\n",
    "金融领域中，收益（return）通常指的是某资产价格的百分比变化"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Date\n",
       "2012-07-23    74.676285\n",
       "2012-07-24    74.316422\n",
       "2012-07-25    71.107155\n",
       "2012-07-26    71.096016\n",
       "2012-07-27    72.367340\n",
       "Name: Adj Close, dtype: float64"
      ]
     },
     "execution_count": 111,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 2011到2012年间苹果公司的股票价格数据\n",
    "import  pandas_datareader.data as web\n",
    "price = web.get_data_yahoo('AAPL','2011-01-01','2012-07-27')['Adj Close']\n",
    "price[-5:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.07239946865659364"
      ]
     },
     "execution_count": 114,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 计算两个时间点指甲的累计百分比回报只需计算价格的百分比变化即可\n",
    "price['2011-10-03']/price['2011-3-01']-1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Date\n",
       "2010-12-31         NaN\n",
       "2011-01-03    0.021732\n",
       "2011-01-04    0.005219\n",
       "2011-01-05    0.008180\n",
       "2011-01-06   -0.000808\n",
       "2011-01-07    0.007161\n",
       "2011-01-10    0.018833\n",
       "2011-01-11   -0.002365\n",
       "2011-01-12    0.008137\n",
       "2011-01-13    0.003658\n",
       "2011-01-14    0.008100\n",
       "2011-01-18   -0.022469\n",
       "2011-01-19   -0.005313\n",
       "2011-01-20   -0.018180\n",
       "2011-01-21   -0.017915\n",
       "2011-01-24    0.032842\n",
       "2011-01-25    0.011706\n",
       "2011-01-26    0.007176\n",
       "2011-01-27   -0.001861\n",
       "2011-01-28   -0.020716\n",
       "2011-01-31    0.009581\n",
       "2011-02-01    0.016827\n",
       "2011-02-02   -0.002058\n",
       "2011-02-03   -0.002556\n",
       "2011-02-04    0.008910\n",
       "2011-02-07    0.015527\n",
       "2011-02-08    0.009435\n",
       "2011-02-09    0.008333\n",
       "2011-02-10   -0.010107\n",
       "2011-02-11    0.006516\n",
       "                ...   \n",
       "2012-06-15    0.004549\n",
       "2012-06-18    0.020291\n",
       "2012-06-19    0.002783\n",
       "2012-06-20   -0.002843\n",
       "2012-06-21   -0.013777\n",
       "2012-06-22    0.007669\n",
       "2012-06-25   -0.019464\n",
       "2012-06-26    0.002207\n",
       "2012-06-27    0.004318\n",
       "2012-06-28   -0.009487\n",
       "2012-06-29    0.026272\n",
       "2012-07-02    0.014589\n",
       "2012-07-03    0.011628\n",
       "2012-07-05    0.017567\n",
       "2012-07-06   -0.006656\n",
       "2012-07-09    0.013220\n",
       "2012-07-10   -0.009252\n",
       "2012-07-11   -0.006215\n",
       "2012-07-12   -0.009150\n",
       "2012-07-13    0.010135\n",
       "2012-07-16    0.003207\n",
       "2012-07-17    0.000049\n",
       "2012-07-18   -0.001120\n",
       "2012-07-19    0.013295\n",
       "2012-07-20   -0.016311\n",
       "2012-07-23   -0.000778\n",
       "2012-07-24   -0.004819\n",
       "2012-07-25   -0.043184\n",
       "2012-07-26   -0.000157\n",
       "2012-07-27    0.017882\n",
       "Name: Adj Close, Length: 397, dtype: float64"
      ]
     },
     "execution_count": 115,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 收益指数是一个表示单位投资（比如1美元）收益的时间序列，从收益指数中可以得出许多假设，比如决定是否进行利润再投资\n",
    "# 通过cumprod计算出一个简单的收益指数\n",
    "returns = price.pct_change()  # 表示当前元素与先前元素相差的百分比，这里即当前元素-前一个元素\n",
    "returns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Date\n",
       "2010-12-31    1.000000\n",
       "2011-01-03    1.021732\n",
       "2011-01-04    1.027065\n",
       "2011-01-05    1.035466\n",
       "2011-01-06    1.034629\n",
       "2011-01-07    1.042038\n",
       "2011-01-10    1.061663\n",
       "2011-01-11    1.059152\n",
       "2011-01-12    1.067770\n",
       "2011-01-13    1.071676\n",
       "2011-01-14    1.080357\n",
       "2011-01-18    1.056083\n",
       "2011-01-19    1.050471\n",
       "2011-01-20    1.031374\n",
       "2011-01-21    1.012897\n",
       "2011-01-24    1.046162\n",
       "2011-01-25    1.058408\n",
       "2011-01-26    1.066003\n",
       "2011-01-27    1.064019\n",
       "2011-01-28    1.041977\n",
       "2011-01-31    1.051959\n",
       "2011-02-01    1.069661\n",
       "2011-02-02    1.067460\n",
       "2011-02-03    1.064732\n",
       "2011-02-04    1.074218\n",
       "2011-02-07    1.090897\n",
       "2011-02-08    1.101190\n",
       "2011-02-09    1.110367\n",
       "2011-02-10    1.099144\n",
       "2011-02-11    1.106306\n",
       "                ...   \n",
       "2012-06-15    1.779916\n",
       "2012-06-18    1.816034\n",
       "2012-06-19    1.821088\n",
       "2012-06-20    1.815910\n",
       "2012-06-21    1.790891\n",
       "2012-06-22    1.804625\n",
       "2012-06-25    1.769500\n",
       "2012-06-26    1.773406\n",
       "2012-06-27    1.781064\n",
       "2012-06-28    1.764167\n",
       "2012-06-29    1.810516\n",
       "2012-07-02    1.836929\n",
       "2012-07-03    1.858289\n",
       "2012-07-05    1.890934\n",
       "2012-07-06    1.878348\n",
       "2012-07-09    1.903181\n",
       "2012-07-10    1.885572\n",
       "2012-07-11    1.873853\n",
       "2012-07-12    1.856708\n",
       "2012-07-13    1.875526\n",
       "2012-07-16    1.881541\n",
       "2012-07-17    1.881634\n",
       "2012-07-18    1.879526\n",
       "2012-07-19    1.904514\n",
       "2012-07-20    1.873450\n",
       "2012-07-23    1.871992\n",
       "2012-07-24    1.862971\n",
       "2012-07-25    1.782521\n",
       "2012-07-26    1.782242\n",
       "2012-07-27    1.814111\n",
       "Name: Adj Close, Length: 397, dtype: float64"
      ]
     },
     "execution_count": 122,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ret_index = (1+returns).cumprod()   # 累积乘积\n",
    "ret_index[0] = 1 # 将第一个值设置为1\n",
    "ret_index"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "D:\\Python\\anaconda\\lib\\site-packages\\ipykernel_launcher.py:2: FutureWarning: how in .resample() is deprecated\n",
      "the new syntax is .resample(...).last()\n",
      "  \n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Date\n",
       "2012-01-31    0.127111\n",
       "2012-02-29    0.188310\n",
       "2012-03-30    0.105284\n",
       "2012-04-30   -0.025970\n",
       "2012-05-31   -0.010703\n",
       "2012-06-29    0.010853\n",
       "2012-07-31    0.001986\n",
       "Freq: BM, Name: Adj Close, dtype: float64"
      ]
     },
     "execution_count": 124,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 得到收益指数之后，计算指定时期内的累计收益\n",
    "m_returns = ret_index.resample('BM',how='last').pct_change()\n",
    "m_returns['2012']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "D:\\Python\\anaconda\\lib\\site-packages\\ipykernel_launcher.py:2: FutureWarning: how in .resample() is deprecated\n",
      "the new syntax is .resample(...).prod()\n",
      "  \n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Date\n",
       "2012-01    0.127111\n",
       "2012-02    0.188310\n",
       "2012-03    0.105284\n",
       "2012-04   -0.025970\n",
       "2012-05   -0.010703\n",
       "2012-06    0.010853\n",
       "2012-07    0.001986\n",
       "Freq: M, Name: Adj Close, dtype: float64"
      ]
     },
     "execution_count": 125,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 上面的结果也能通过重采样聚类从日百分比变化中计算得出\n",
    "m_rets = (1+returns).resample(\"M\",how='prod',kind='period')-1\n",
    "m_rets['2012']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 分组变换和分析"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['MYNBI', 'QPMZJ', 'PLSGQ', 'EJEYD', 'TZIRW'], dtype='<U5')"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 以一组假想的股票投资组合为例。首先随机生成1000个股票代码\n",
    "import random\n",
    "import string\n",
    "random.seed(0)\n",
    "N = 1000\n",
    "def rands(n):\n",
    "    choices = string.ascii_uppercase\n",
    "    return ''.join([random.choice(choices) for _ in range(n)])\n",
    "tickers = np.array([rands(5) for _ in range(N)])  # 随机生成含5个大写字母的代码\n",
    "tickers[:5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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>Momentum</th>\n",
       "      <th>ShortInterest</th>\n",
       "      <th>Value</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>MYNBI</th>\n",
       "      <td>0.033735</td>\n",
       "      <td>-0.016758</td>\n",
       "      <td>0.070012</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>QPMZJ</th>\n",
       "      <td>0.039123</td>\n",
       "      <td>-0.021287</td>\n",
       "      <td>0.092283</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PLSGQ</th>\n",
       "      <td>0.019385</td>\n",
       "      <td>-0.021671</td>\n",
       "      <td>0.074507</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>EJEYD</th>\n",
       "      <td>0.028480</td>\n",
       "      <td>-0.014490</td>\n",
       "      <td>0.084253</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TZIRW</th>\n",
       "      <td>0.026526</td>\n",
       "      <td>-0.022802</td>\n",
       "      <td>0.077527</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       Momentum  ShortInterest     Value\n",
       "MYNBI  0.033735      -0.016758  0.070012\n",
       "QPMZJ  0.039123      -0.021287  0.092283\n",
       "PLSGQ  0.019385      -0.021671  0.074507\n",
       "EJEYD  0.028480      -0.014490  0.084253\n",
       "TZIRW  0.026526      -0.022802  0.077527"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 创建一个含有3列的Dataframe来承载这些假想数据，不过只选择部分股票组成改投资组合\n",
    "M =  500\n",
    "df = pd.DataFrame({'Momentum':np.random.randn(M)/200 + 0.03,\n",
    "                   'Value':np.random.randn(M)/200 + 0.08,\n",
    "                   'ShortInterest':np.random.randn(M)/200 -0.02},\n",
    "                  index = tickers[:M])\n",
    "df[:5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MYNBI    FINANCIAL\n",
       "QPMZJ    FINANCIAL\n",
       "PLSGQ         TECH\n",
       "EJEYD    FINANCIAL\n",
       "TZIRW         TECH\n",
       "Name: industry, dtype: object"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 为这些股票随机创建一个行业分类,为简单起见，只选两个行业，并将映射关系保存在Series中\n",
    "ind_names = np.array(['FINANCIAL','TECH'])\n",
    "sampler = np.random.randint(0,len(ind_names),N) # 生成N个[0，len(ind_names)）的随机整数\n",
    "industries = pd.Series(ind_names[sampler],index = tickers,name='industry')\n",
    "industries[:5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "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>Momentum</th>\n",
       "      <th>ShortInterest</th>\n",
       "      <th>Value</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>industry</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>FINANCIAL</th>\n",
       "      <td>0.030148</td>\n",
       "      <td>-0.020076</td>\n",
       "      <td>0.080253</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TECH</th>\n",
       "      <td>0.029665</td>\n",
       "      <td>-0.019782</td>\n",
       "      <td>0.079828</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           Momentum  ShortInterest     Value\n",
       "industry                                    \n",
       "FINANCIAL  0.030148      -0.020076  0.080253\n",
       "TECH       0.029665      -0.019782  0.079828"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 根据行业分类进行分组并执行分组聚合和变换\n",
    "by_industry = df.groupby(industries)\n",
    "by_industry.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "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=\"8\" halign=\"left\">Momentum</th>\n",
       "      <th colspan=\"5\" halign=\"left\">ShortInterest</th>\n",
       "      <th colspan=\"8\" halign=\"left\">Value</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>min</th>\n",
       "      <th>25%</th>\n",
       "      <th>50%</th>\n",
       "      <th>75%</th>\n",
       "      <th>max</th>\n",
       "      <th>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>...</th>\n",
       "      <th>75%</th>\n",
       "      <th>max</th>\n",
       "      <th>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>min</th>\n",
       "      <th>25%</th>\n",
       "      <th>50%</th>\n",
       "      <th>75%</th>\n",
       "      <th>max</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>industry</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>FINANCIAL</th>\n",
       "      <td>247.0</td>\n",
       "      <td>0.030148</td>\n",
       "      <td>0.004984</td>\n",
       "      <td>0.016261</td>\n",
       "      <td>0.026629</td>\n",
       "      <td>0.029981</td>\n",
       "      <td>0.032885</td>\n",
       "      <td>0.045733</td>\n",
       "      <td>247.0</td>\n",
       "      <td>-0.020076</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.016260</td>\n",
       "      <td>-0.006750</td>\n",
       "      <td>247.0</td>\n",
       "      <td>0.080253</td>\n",
       "      <td>0.005368</td>\n",
       "      <td>0.061128</td>\n",
       "      <td>0.076654</td>\n",
       "      <td>0.080299</td>\n",
       "      <td>0.083426</td>\n",
       "      <td>0.095904</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TECH</th>\n",
       "      <td>253.0</td>\n",
       "      <td>0.029665</td>\n",
       "      <td>0.004807</td>\n",
       "      <td>0.016640</td>\n",
       "      <td>0.026401</td>\n",
       "      <td>0.029609</td>\n",
       "      <td>0.032789</td>\n",
       "      <td>0.043148</td>\n",
       "      <td>253.0</td>\n",
       "      <td>-0.019782</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.016611</td>\n",
       "      <td>-0.005772</td>\n",
       "      <td>253.0</td>\n",
       "      <td>0.079828</td>\n",
       "      <td>0.004884</td>\n",
       "      <td>0.063029</td>\n",
       "      <td>0.076666</td>\n",
       "      <td>0.080368</td>\n",
       "      <td>0.083261</td>\n",
       "      <td>0.092368</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2 rows × 24 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          Momentum                                                    \\\n",
       "             count      mean       std       min       25%       50%   \n",
       "industry                                                               \n",
       "FINANCIAL    247.0  0.030148  0.004984  0.016261  0.026629  0.029981   \n",
       "TECH         253.0  0.029665  0.004807  0.016640  0.026401  0.029609   \n",
       "\n",
       "                              ShortInterest              ...               \\\n",
       "                75%       max         count      mean    ...          75%   \n",
       "industry                                                 ...                \n",
       "FINANCIAL  0.032885  0.045733         247.0 -0.020076    ...    -0.016260   \n",
       "TECH       0.032789  0.043148         253.0 -0.019782    ...    -0.016611   \n",
       "\n",
       "                     Value                                                    \\\n",
       "                max  count      mean       std       min       25%       50%   \n",
       "industry                                                                       \n",
       "FINANCIAL -0.006750  247.0  0.080253  0.005368  0.061128  0.076654  0.080299   \n",
       "TECH      -0.005772  253.0  0.079828  0.004884  0.063029  0.076666  0.080368   \n",
       "\n",
       "                               \n",
       "                75%       max  \n",
       "industry                       \n",
       "FINANCIAL  0.083426  0.095904  \n",
       "TECH       0.083261  0.092368  \n",
       "\n",
       "[2 rows x 24 columns]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "by_industry.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "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>Momentum</th>\n",
       "      <th>ShortInterest</th>\n",
       "      <th>Value</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>MYNBI</th>\n",
       "      <td>0.719760</td>\n",
       "      <td>0.636247</td>\n",
       "      <td>-1.907817</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>QPMZJ</th>\n",
       "      <td>1.800712</td>\n",
       "      <td>-0.232401</td>\n",
       "      <td>2.240880</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PLSGQ</th>\n",
       "      <td>-2.138416</td>\n",
       "      <td>-0.393543</td>\n",
       "      <td>-1.089380</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>EJEYD</th>\n",
       "      <td>-0.334460</td>\n",
       "      <td>1.071347</td>\n",
       "      <td>0.745003</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TZIRW</th>\n",
       "      <td>-0.652991</td>\n",
       "      <td>-0.629134</td>\n",
       "      <td>-0.471077</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       Momentum  ShortInterest     Value\n",
       "MYNBI  0.719760       0.636247 -1.907817\n",
       "QPMZJ  1.800712      -0.232401  2.240880\n",
       "PLSGQ -2.138416      -0.393543 -1.089380\n",
       "EJEYD -0.334460       1.071347  0.745003\n",
       "TZIRW -0.652991      -0.629134 -0.471077"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 对这些按行业分组的投资组合进行各种变换，可以白哪些自定义的变化函数，如行业内标准化处理，它广泛用于股票投资组合的构建过程\n",
    "# 行业内标准化处理\n",
    "def zscore(group):\n",
    "    return (group - group.mean())/group.std()\n",
    "df_stand = by_industry.apply(zscore)\n",
    "df_stand.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "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\">Momentum</th>\n",
       "      <th colspan=\"2\" halign=\"left\">ShortInterest</th>\n",
       "      <th colspan=\"2\" halign=\"left\">Value</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>industry</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>FINANCIAL</th>\n",
       "      <td>-6.679317e-16</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.017280e-15</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-1.286061e-14</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TECH</th>\n",
       "      <td>2.450389e-15</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.660664e-15</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-3.783096e-15</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               Momentum      ShortInterest              Value     \n",
       "                   mean  std          mean  std          mean  std\n",
       "industry                                                          \n",
       "FINANCIAL -6.679317e-16  1.0  2.017280e-15  1.0 -1.286061e-14  1.0\n",
       "TECH       2.450389e-15  1.0  3.660664e-15  1.0 -3.783096e-15  1.0"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 这样处理后，各行业的平均值为0，标准差为1\n",
    "df_stand.groupby(industries).agg(['mean','std'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "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\">Momentum</th>\n",
       "      <th colspan=\"2\" halign=\"left\">ShortInterest</th>\n",
       "      <th colspan=\"2\" halign=\"left\">Value</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>min</th>\n",
       "      <th>max</th>\n",
       "      <th>min</th>\n",
       "      <th>max</th>\n",
       "      <th>min</th>\n",
       "      <th>max</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>industry</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>FINANCIAL</th>\n",
       "      <td>1.0</td>\n",
       "      <td>247.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>247.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>247.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TECH</th>\n",
       "      <td>1.0</td>\n",
       "      <td>253.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>253.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>253.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          Momentum        ShortInterest        Value       \n",
       "               min    max           min    max   min    max\n",
       "industry                                                   \n",
       "FINANCIAL      1.0  247.0           1.0  247.0   1.0  247.0\n",
       "TECH           1.0  253.0           1.0  253.0   1.0  253.0"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 内置变化函数（如rank）的用法更简洁一些\n",
    "# 行业内降序排名\n",
    "ind_rank = by_industry.rank(ascending=False)\n",
    "ind_rank.groupby(industries).agg(['min','max'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "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>Momentum</th>\n",
       "      <th>ShortInterest</th>\n",
       "      <th>Value</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>MYNBI</th>\n",
       "      <td>0.951755</td>\n",
       "      <td>0.811791</td>\n",
       "      <td>-1.665571</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>QPMZJ</th>\n",
       "      <td>1.553600</td>\n",
       "      <td>-0.265932</td>\n",
       "      <td>1.679568</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PLSGQ</th>\n",
       "      <td>-1.639815</td>\n",
       "      <td>-0.505610</td>\n",
       "      <td>-1.311852</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>EJEYD</th>\n",
       "      <td>-0.419892</td>\n",
       "      <td>1.231683</td>\n",
       "      <td>0.965752</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TZIRW</th>\n",
       "      <td>-0.819907</td>\n",
       "      <td>-0.833572</td>\n",
       "      <td>-0.683256</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ZTEJD</th>\n",
       "      <td>-0.097975</td>\n",
       "      <td>-1.693564</td>\n",
       "      <td>0.559856</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>XCVKP</th>\n",
       "      <td>1.413636</td>\n",
       "      <td>0.321917</td>\n",
       "      <td>1.035734</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RDLNK</th>\n",
       "      <td>-0.983889</td>\n",
       "      <td>0.355293</td>\n",
       "      <td>-1.653480</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TUGRP</th>\n",
       "      <td>0.054660</td>\n",
       "      <td>0.628596</td>\n",
       "      <td>-1.421173</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>OQIBZ</th>\n",
       "      <td>1.229861</td>\n",
       "      <td>1.503163</td>\n",
       "      <td>0.751582</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RACXM</th>\n",
       "      <td>0.273302</td>\n",
       "      <td>1.147870</td>\n",
       "      <td>0.409954</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>WZVUA</th>\n",
       "      <td>0.040995</td>\n",
       "      <td>-1.694475</td>\n",
       "      <td>-0.218642</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TPKHX</th>\n",
       "      <td>1.287669</td>\n",
       "      <td>-0.027993</td>\n",
       "      <td>-0.503870</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>KWCGS</th>\n",
       "      <td>-0.405896</td>\n",
       "      <td>0.713816</td>\n",
       "      <td>0.643834</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HHZEZ</th>\n",
       "      <td>0.545860</td>\n",
       "      <td>-0.909766</td>\n",
       "      <td>-0.363906</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ROCCK</th>\n",
       "      <td>1.245679</td>\n",
       "      <td>1.469622</td>\n",
       "      <td>0.979748</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>QPDJR</th>\n",
       "      <td>-0.783798</td>\n",
       "      <td>0.853780</td>\n",
       "      <td>0.055986</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>JWDRK</th>\n",
       "      <td>-1.455625</td>\n",
       "      <td>-0.419892</td>\n",
       "      <td>-0.419892</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RGZTR</th>\n",
       "      <td>1.216196</td>\n",
       "      <td>1.626149</td>\n",
       "      <td>1.038549</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SJOCT</th>\n",
       "      <td>-1.049730</td>\n",
       "      <td>1.637579</td>\n",
       "      <td>-0.909766</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ZMKSH</th>\n",
       "      <td>0.237939</td>\n",
       "      <td>0.223942</td>\n",
       "      <td>-0.937759</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>JFGFB</th>\n",
       "      <td>-1.339182</td>\n",
       "      <td>-1.188866</td>\n",
       "      <td>1.694475</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TVIPC</th>\n",
       "      <td>-1.581593</td>\n",
       "      <td>-1.343654</td>\n",
       "      <td>1.413636</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CVYEE</th>\n",
       "      <td>0.204977</td>\n",
       "      <td>-0.587600</td>\n",
       "      <td>0.232307</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BCWRV</th>\n",
       "      <td>1.455625</td>\n",
       "      <td>-1.091719</td>\n",
       "      <td>1.707561</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MWQIQ</th>\n",
       "      <td>-0.811791</td>\n",
       "      <td>-1.679568</td>\n",
       "      <td>0.405896</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ZHGVS</th>\n",
       "      <td>-0.792577</td>\n",
       "      <td>0.833572</td>\n",
       "      <td>-1.284521</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NSIOP</th>\n",
       "      <td>1.311852</td>\n",
       "      <td>1.284521</td>\n",
       "      <td>-1.516829</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>VUWZL</th>\n",
       "      <td>-0.942893</td>\n",
       "      <td>1.325517</td>\n",
       "      <td>1.298187</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CKTDP</th>\n",
       "      <td>-1.007741</td>\n",
       "      <td>0.951755</td>\n",
       "      <td>-1.063726</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CCPMS</th>\n",
       "      <td>0.997554</td>\n",
       "      <td>-1.339182</td>\n",
       "      <td>-0.273302</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NRCIP</th>\n",
       "      <td>-1.161535</td>\n",
       "      <td>-0.997554</td>\n",
       "      <td>-1.544159</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HDJEL</th>\n",
       "      <td>0.629838</td>\n",
       "      <td>1.161701</td>\n",
       "      <td>-0.881773</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ZDEVB</th>\n",
       "      <td>-1.708140</td>\n",
       "      <td>1.462168</td>\n",
       "      <td>0.942893</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>VESRG</th>\n",
       "      <td>0.083978</td>\n",
       "      <td>-0.279928</td>\n",
       "      <td>-0.615842</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ABMRW</th>\n",
       "      <td>-1.448503</td>\n",
       "      <td>-1.161535</td>\n",
       "      <td>-1.339182</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>YTPWW</th>\n",
       "      <td>-1.626149</td>\n",
       "      <td>1.161535</td>\n",
       "      <td>0.601265</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DPRLY</th>\n",
       "      <td>-0.737917</td>\n",
       "      <td>0.970224</td>\n",
       "      <td>0.191312</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>KDVAH</th>\n",
       "      <td>-0.265932</td>\n",
       "      <td>0.923762</td>\n",
       "      <td>-1.091719</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HPJIH</th>\n",
       "      <td>1.721557</td>\n",
       "      <td>-1.259676</td>\n",
       "      <td>0.727813</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>APLQK</th>\n",
       "      <td>-1.105715</td>\n",
       "      <td>0.069982</td>\n",
       "      <td>0.811791</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CCJSN</th>\n",
       "      <td>-0.054660</td>\n",
       "      <td>0.218642</td>\n",
       "      <td>0.860903</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HXLMY</th>\n",
       "      <td>1.595589</td>\n",
       "      <td>0.587849</td>\n",
       "      <td>-0.349910</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>EHJGX</th>\n",
       "      <td>-0.368958</td>\n",
       "      <td>-1.079545</td>\n",
       "      <td>-1.571489</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>YPVZL</th>\n",
       "      <td>0.055986</td>\n",
       "      <td>-1.497615</td>\n",
       "      <td>-0.993744</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>JMTEY</th>\n",
       "      <td>-0.475878</td>\n",
       "      <td>0.559856</td>\n",
       "      <td>0.363906</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DMLQP</th>\n",
       "      <td>-1.516829</td>\n",
       "      <td>0.601265</td>\n",
       "      <td>0.259637</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HUWLU</th>\n",
       "      <td>0.601845</td>\n",
       "      <td>-0.139964</td>\n",
       "      <td>0.167957</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LNILM</th>\n",
       "      <td>0.819907</td>\n",
       "      <td>-0.136651</td>\n",
       "      <td>0.150316</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>YYWJD</th>\n",
       "      <td>0.825788</td>\n",
       "      <td>0.293924</td>\n",
       "      <td>-0.125968</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PJDOE</th>\n",
       "      <td>1.462168</td>\n",
       "      <td>-0.081991</td>\n",
       "      <td>-0.341628</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LHXFK</th>\n",
       "      <td>1.106875</td>\n",
       "      <td>-0.614930</td>\n",
       "      <td>-0.314298</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PHDVM</th>\n",
       "      <td>1.049730</td>\n",
       "      <td>0.083978</td>\n",
       "      <td>-0.923762</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MOQOS</th>\n",
       "      <td>0.251935</td>\n",
       "      <td>-1.609586</td>\n",
       "      <td>-0.139964</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>THVMQ</th>\n",
       "      <td>0.874568</td>\n",
       "      <td>0.642261</td>\n",
       "      <td>1.216196</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>JPHKQ</th>\n",
       "      <td>0.671827</td>\n",
       "      <td>-0.559856</td>\n",
       "      <td>0.867777</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>VACPK</th>\n",
       "      <td>1.257191</td>\n",
       "      <td>-0.655926</td>\n",
       "      <td>1.065879</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MHNBS</th>\n",
       "      <td>1.371647</td>\n",
       "      <td>1.483618</td>\n",
       "      <td>-0.671827</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>YBNCI</th>\n",
       "      <td>-0.710586</td>\n",
       "      <td>-0.218642</td>\n",
       "      <td>0.013665</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>GXKFD</th>\n",
       "      <td>-1.667145</td>\n",
       "      <td>-0.956559</td>\n",
       "      <td>1.202531</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>500 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       Momentum  ShortInterest     Value\n",
       "MYNBI  0.951755       0.811791 -1.665571\n",
       "QPMZJ  1.553600      -0.265932  1.679568\n",
       "PLSGQ -1.639815      -0.505610 -1.311852\n",
       "EJEYD -0.419892       1.231683  0.965752\n",
       "TZIRW -0.819907      -0.833572 -0.683256\n",
       "ZTEJD -0.097975      -1.693564  0.559856\n",
       "XCVKP  1.413636       0.321917  1.035734\n",
       "RDLNK -0.983889       0.355293 -1.653480\n",
       "TUGRP  0.054660       0.628596 -1.421173\n",
       "OQIBZ  1.229861       1.503163  0.751582\n",
       "RACXM  0.273302       1.147870  0.409954\n",
       "WZVUA  0.040995      -1.694475 -0.218642\n",
       "TPKHX  1.287669      -0.027993 -0.503870\n",
       "KWCGS -0.405896       0.713816  0.643834\n",
       "HHZEZ  0.545860      -0.909766 -0.363906\n",
       "ROCCK  1.245679       1.469622  0.979748\n",
       "QPDJR -0.783798       0.853780  0.055986\n",
       "JWDRK -1.455625      -0.419892 -0.419892\n",
       "RGZTR  1.216196       1.626149  1.038549\n",
       "SJOCT -1.049730       1.637579 -0.909766\n",
       "ZMKSH  0.237939       0.223942 -0.937759\n",
       "JFGFB -1.339182      -1.188866  1.694475\n",
       "TVIPC -1.581593      -1.343654  1.413636\n",
       "CVYEE  0.204977      -0.587600  0.232307\n",
       "BCWRV  1.455625      -1.091719  1.707561\n",
       "MWQIQ -0.811791      -1.679568  0.405896\n",
       "ZHGVS -0.792577       0.833572 -1.284521\n",
       "NSIOP  1.311852       1.284521 -1.516829\n",
       "VUWZL -0.942893       1.325517  1.298187\n",
       "CKTDP -1.007741       0.951755 -1.063726\n",
       "...         ...            ...       ...\n",
       "CCPMS  0.997554      -1.339182 -0.273302\n",
       "NRCIP -1.161535      -0.997554 -1.544159\n",
       "HDJEL  0.629838       1.161701 -0.881773\n",
       "ZDEVB -1.708140       1.462168  0.942893\n",
       "VESRG  0.083978      -0.279928 -0.615842\n",
       "ABMRW -1.448503      -1.161535 -1.339182\n",
       "YTPWW -1.626149       1.161535  0.601265\n",
       "DPRLY -0.737917       0.970224  0.191312\n",
       "KDVAH -0.265932       0.923762 -1.091719\n",
       "HPJIH  1.721557      -1.259676  0.727813\n",
       "APLQK -1.105715       0.069982  0.811791\n",
       "CCJSN -0.054660       0.218642  0.860903\n",
       "HXLMY  1.595589       0.587849 -0.349910\n",
       "EHJGX -0.368958      -1.079545 -1.571489\n",
       "YPVZL  0.055986      -1.497615 -0.993744\n",
       "JMTEY -0.475878       0.559856  0.363906\n",
       "DMLQP -1.516829       0.601265  0.259637\n",
       "HUWLU  0.601845      -0.139964  0.167957\n",
       "LNILM  0.819907      -0.136651  0.150316\n",
       "YYWJD  0.825788       0.293924 -0.125968\n",
       "PJDOE  1.462168      -0.081991 -0.341628\n",
       "LHXFK  1.106875      -0.614930 -0.314298\n",
       "PHDVM  1.049730       0.083978 -0.923762\n",
       "MOQOS  0.251935      -1.609586 -0.139964\n",
       "THVMQ  0.874568       0.642261  1.216196\n",
       "JPHKQ  0.671827      -0.559856  0.867777\n",
       "VACPK  1.257191      -0.655926  1.065879\n",
       "MHNBS  1.371647       1.483618 -0.671827\n",
       "YBNCI -0.710586      -0.218642  0.013665\n",
       "GXKFD -1.667145      -0.956559  1.202531\n",
       "\n",
       "[500 rows x 3 columns]"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 在股票投资组合的定量分析中，“排名和标准化”是一种很常见的变换运算组合。\n",
    "# 通过将rank和zscore链接在一起即可完成整个变换过程\n",
    "# 行业内排名和标准化\n",
    "by_industry.apply(lambda x: zscore(x.rank()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 分组因子暴露\n",
    "因子分析是投资组合定量管理中的一个技术。投资组合的持有量和性能（收益与损失）可以被分解为一个或多个表示投资组合权重的因子（风险因子就是其中之一）。\n",
    "- 某只股票的价格与某个基准（比如标准普尔500指数）的协动性被称为其贝塔风险系数（beta,一种常见的风险因子）"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['UNUJV', 'HMGTW', 'FJMOJ', 'ETQCI', 'PMENZ'], dtype='<U5')"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 以一个人为构成的投资组合为例，它由三个随机生成的因子（通常称为因子载荷）和一些权重构成\n",
    "from numpy.random import rand\n",
    "import numpy as np\n",
    "fac1,fac2,fac3 = np.random.rand(3,1000)\n",
    "ticker_subset = tickers.take(np.random.permutation(N))  # permutaion(N):对[0,N)之间的序列进行随机排序\n",
    "ticker_subset[:5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "UNUJV    0.968903\n",
       "HMGTW   -0.246789\n",
       "FJMOJ    1.047932\n",
       "ETQCI    1.528396\n",
       "PMENZ   -0.180584\n",
       "dtype: float64"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 因子加权和以及噪声\n",
    "port = pd.Series(0.7*fac1-1.2*fac2+0.3*fac3+rand(1000),\n",
    "                index = ticker_subset)\n",
    "port[:5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "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>f1</th>\n",
       "      <th>f2</th>\n",
       "      <th>f3</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>UNUJV</th>\n",
       "      <td>0.310559</td>\n",
       "      <td>0.153851</td>\n",
       "      <td>0.908788</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HMGTW</th>\n",
       "      <td>0.101951</td>\n",
       "      <td>0.882656</td>\n",
       "      <td>0.774865</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FJMOJ</th>\n",
       "      <td>0.972846</td>\n",
       "      <td>0.500346</td>\n",
       "      <td>0.265660</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ETQCI</th>\n",
       "      <td>0.854082</td>\n",
       "      <td>0.168083</td>\n",
       "      <td>0.745699</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PMENZ</th>\n",
       "      <td>0.466477</td>\n",
       "      <td>0.713596</td>\n",
       "      <td>0.434021</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             f1        f2        f3\n",
       "UNUJV  0.310559  0.153851  0.908788\n",
       "HMGTW  0.101951  0.882656  0.774865\n",
       "FJMOJ  0.972846  0.500346  0.265660\n",
       "ETQCI  0.854082  0.168083  0.745699\n",
       "PMENZ  0.466477  0.713596  0.434021"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "factors = pd.DataFrame({'f1':fac1,'f2':fac2,'f3':fac3},\n",
    "                       index=ticker_subset)\n",
    "factors[:5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "f1    0.398062\n",
       "f2   -0.699874\n",
       "f3    0.119866\n",
       "dtype: float64"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 各因子与投资组合之间的矢量相关性可能说明不了什么问题\n",
    "factors.corrwith(port)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "const    0.544551\n",
      "f1       0.723653\n",
      "f2      -1.229669\n",
      "f3       0.230682\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "# 计算因子暴露的标准方式是最小二乘回归，使用statsmodels.api.ols(将factors作为解释变量)即可计算出整个投资组合的暴露\n",
    "import statsmodels.api as sm\n",
    "X = sm.add_constant(factors)  # 增加常数项\n",
    "model = sm.OLS(port,X)\n",
    "results = model.fit()\n",
    "print(results.params)  # 即系数beta值"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:                      y   R-squared:                       0.676\n",
      "Model:                            OLS   Adj. R-squared:                  0.675\n",
      "Method:                 Least Squares   F-statistic:                     693.9\n",
      "Date:                Tue, 23 Jun 2020   Prob (F-statistic):          2.01e-243\n",
      "Time:                        21:44:32   Log-Likelihood:                -183.76\n",
      "No. Observations:                1000   AIC:                             375.5\n",
      "Df Residuals:                     996   BIC:                             395.2\n",
      "Df Model:                           3                                         \n",
      "Covariance Type:            nonrobust                                         \n",
      "==============================================================================\n",
      "                 coef    std err          t      P>|t|      [0.025      0.975]\n",
      "------------------------------------------------------------------------------\n",
      "const          0.5446      0.029     18.939      0.000       0.488       0.601\n",
      "f1             0.7237      0.032     22.786      0.000       0.661       0.786\n",
      "f2            -1.2297      0.031    -39.392      0.000      -1.291      -1.168\n",
      "f3             0.2307      0.032      7.281      0.000       0.169       0.293\n",
      "==============================================================================\n",
      "Omnibus:                      820.210   Durbin-Watson:                   2.033\n",
      "Prob(Omnibus):                  0.000   Jarque-Bera (JB):               61.483\n",
      "Skew:                          -0.029   Prob(JB):                     4.46e-14\n",
      "Kurtosis:                       1.787   Cond. No.                         6.10\n",
      "==============================================================================\n",
      "\n",
      "Warnings:\n",
      "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n"
     ]
    }
   ],
   "source": [
    "print(results.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "FINANCIAL的暴露量\n",
      "const    0.532183\n",
      "f1       0.742246\n",
      "f2      -1.269758\n",
      "f3       0.278664\n",
      "dtype: float64\n",
      "TECH的暴露量\n",
      "const    0.552136\n",
      "f1       0.707606\n",
      "f2      -1.188736\n",
      "f3       0.187569\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "# 根据行业进行分组，并应用上面的函数，传入因子载荷的DataFrame\n",
    "by_ind = port.groupby(industries)\n",
    "\n",
    "# 可以通过groupby计算各行业的暴露量\n",
    "def CalExpobyInd(by_ind):\n",
    "    groupedItems=[]\n",
    "    for i in by_ind:\n",
    "        groupedItems.append(i)\n",
    "   \n",
    "    for i in range(2):\n",
    "        y=s[i][1]\n",
    "        indexs=s[i][1].keys().tolist()\n",
    "        x=factors.ix[indexs]\n",
    "        grouped = y.agg(beta_exposure,factors=x)\n",
    "        print(s[i][0]+'的暴露量')\n",
    "        print(grouped)\n",
    "\n",
    "def beta_exposure(chunk,factors=None):\n",
    "    X = sm.add_constant(factors)  # 增加常数项\n",
    "    model = sm.OLS(chunk,X)\n",
    "    results = model.fit()\n",
    "    return results.params\n",
    "\n",
    "CalExpobyInd(by_ind)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 十分位和四分位的分析\n",
    "股票投资的性能可以根据各股的市盈率被划入四分位（四个大小相等的块）。通过pandas.qcut和groupby可以实现分位数"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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>High</th>\n",
       "      <th>Low</th>\n",
       "      <th>Open</th>\n",
       "      <th>Close</th>\n",
       "      <th>Volume</th>\n",
       "      <th>Adj Close</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</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>2012-07-23</th>\n",
       "      <td>136.380005</td>\n",
       "      <td>133.839996</td>\n",
       "      <td>134.470001</td>\n",
       "      <td>135.089996</td>\n",
       "      <td>145210900.0</td>\n",
       "      <td>115.059578</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-07-24</th>\n",
       "      <td>135.250000</td>\n",
       "      <td>133.029999</td>\n",
       "      <td>135.190002</td>\n",
       "      <td>133.929993</td>\n",
       "      <td>173301200.0</td>\n",
       "      <td>114.071533</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-07-25</th>\n",
       "      <td>134.559998</td>\n",
       "      <td>133.250000</td>\n",
       "      <td>134.210007</td>\n",
       "      <td>133.960007</td>\n",
       "      <td>129122300.0</td>\n",
       "      <td>114.097115</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-07-26</th>\n",
       "      <td>136.460007</td>\n",
       "      <td>135.259995</td>\n",
       "      <td>135.889999</td>\n",
       "      <td>136.169998</td>\n",
       "      <td>156526500.0</td>\n",
       "      <td>115.979378</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-07-27</th>\n",
       "      <td>139.070007</td>\n",
       "      <td>136.139999</td>\n",
       "      <td>136.889999</td>\n",
       "      <td>138.679993</td>\n",
       "      <td>236768900.0</td>\n",
       "      <td>118.117233</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  High         Low        Open       Close       Volume  \\\n",
       "Date                                                                      \n",
       "2012-07-23  136.380005  133.839996  134.470001  135.089996  145210900.0   \n",
       "2012-07-24  135.250000  133.029999  135.190002  133.929993  173301200.0   \n",
       "2012-07-25  134.559998  133.250000  134.210007  133.960007  129122300.0   \n",
       "2012-07-26  136.460007  135.259995  135.889999  136.169998  156526500.0   \n",
       "2012-07-27  139.070007  136.139999  136.889999  138.679993  236768900.0   \n",
       "\n",
       "             Adj Close  \n",
       "Date                    \n",
       "2012-07-23  115.059578  \n",
       "2012-07-24  114.071533  \n",
       "2012-07-25  114.097115  \n",
       "2012-07-26  115.979378  \n",
       "2012-07-27  118.117233  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 利用跟随策略或动量交易策略通过SPX交易所交易基金买卖标准普尔500指数\n",
    "# 下载历史价格\n",
    "import pandas_datareader.data as web\n",
    "data = web.get_data_yahoo(\"SPY\",'2006-01-01','2012-07-27')\n",
    "data.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 222,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Date\n",
       "2006-01-03         NaN\n",
       "2006-01-04    0.004736\n",
       "2006-01-05    0.000628\n",
       "2006-01-06    0.008321\n",
       "2006-01-09    0.002569\n",
       "Name: Adj Close, dtype: float64"
      ]
     },
     "execution_count": 222,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 计算日收益率，并编写一个用于将收益率变化为趋势信号（通过滞后移动形式）的函数\n",
    "px = data['Adj Close']\n",
    "returns = px.pct_change()  # 当前项\n",
    "returns.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 224,
   "metadata": {},
   "outputs": [],
   "source": [
    "def to_index(rets):\n",
    "    index = (1+rets).cumprod()\n",
    "    first_loc = max(index.notnull().argmax()-pd.Timedelta(days=1),index.index[0])\n",
    "    index[first_loc]=1\n",
    "    return index\n",
    "\n",
    "def trend_signal(rets,lookback,lag):\n",
    "    signal=pd.rolling_sum(rets,lookback,min_periods=lookback-5)\n",
    "    return signal.shift(lag)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 225,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "D:\\Python\\anaconda\\lib\\site-packages\\ipykernel_launcher.py:8: FutureWarning: pd.rolling_sum is deprecated for Series and will be removed in a future version, replace with \n",
      "\tSeries.rolling(window=100,min_periods=95,center=False).sum()\n",
      "  \n",
      "D:\\Python\\anaconda\\lib\\site-packages\\ipykernel_launcher.py:3: FutureWarning: \n",
      ".resample() is now a deferred operation\n",
      "You called resample(...) on this deferred object which materialized it into a series\n",
      "by implicitly taking the mean.  Use .resample(...).mean() instead\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n",
      "D:\\Python\\anaconda\\lib\\site-packages\\ipykernel_launcher.py:3: FutureWarning: fill_method is deprecated to .resample()\n",
      "the new syntax is .resample(...).ffill()\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Date\n",
       "2012-07-23    0.004223\n",
       "2012-07-24    0.004223\n",
       "2012-07-25    0.004223\n",
       "2012-07-26    0.004223\n",
       "2012-07-27    0.005242\n",
       "Freq: B, Name: Adj Close, dtype: float64"
      ]
     },
     "execution_count": 225,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 创建和测试一种根据每周五动量信号进行交易的交易策略\n",
    "signal = trend_signal(returns,100,3)\n",
    "trade_friday = signal.resample('W-FRI').resample('B',fill_method='ffill')  # \n",
    "trade_friday.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 226,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Date\n",
       "2012-07-23    0.004223\n",
       "2012-07-24    0.004223\n",
       "2012-07-25    0.004223\n",
       "2012-07-26    0.004223\n",
       "2012-07-27    0.004223\n",
       "Freq: B, Name: Adj Close, dtype: float64"
      ]
     },
     "execution_count": 226,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trade_friday.shift(1).tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 227,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Date\n",
       "2012-07-23   -0.010112\n",
       "2012-07-24   -0.008587\n",
       "2012-07-25    0.000224\n",
       "2012-07-26    0.016497\n",
       "2012-07-27    0.018433\n",
       "Name: Adj Close, dtype: float64"
      ]
     },
     "execution_count": 227,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "returns.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 228,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Date\n",
       "2006-01-03   NaN\n",
       "2006-01-04   NaN\n",
       "2006-01-05   NaN\n",
       "2006-01-06   NaN\n",
       "2006-01-09   NaN\n",
       "Name: Adj Close, dtype: float64"
      ]
     },
     "execution_count": 228,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trade_rets = trade_friday.shift(1)*returns\n",
    "trade_rets.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 229,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "D:\\Python\\anaconda\\lib\\site-packages\\ipykernel_launcher.py:3: FutureWarning: 'argmax' is deprecated. Use 'idxmax' instead. The behavior of 'argmax' will be corrected to return the positional maximum in the future. Use 'series.values.argmax' to get the position of the maximum now.\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0xeed40f0>"
      ]
     },
     "execution_count": 229,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXsAAAEGCAYAAACEgjUUAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvNQv5yAAAIABJREFUeJzt3Xmck9W9+PHPSTI7M8MAw84wLCKbIIIIuLKpFVyuvXXFauuVW6tdbG1/VavV2qu2t9q6tUptbWu1Wq1eV9wAEVRAQDaRfR3WGQZmX7Kc3x/Jkz0zmSyTJ5Pv+/XiNcmT50nOw2S+Ofk+53yP0lojhBCia7OkugFCCCGST4K9EEJkAAn2QgiRASTYCyFEBpBgL4QQGUCCvRBCZAAJ9kIIkQEk2AshRAaQYC+EEBnAluoGGHr16qXLy8tT3QwhhEgra9asqdJal7a3n2mCfXl5OatXr051M4QQIq0opfZGs5+kcYQQIgNIsBdCiAwgwV4IITKABHshhMgAEuyFECIDRBXslVJZSqk3O7KPUipXKfWWUmq9Uuo5pZSKt7FCCCFi026wV0rlAWuA2R3cZx5QobUeD5S0dbwQQojkajfYa62btNbjgIoO7jMD+MBzezEwPZ6GChErrTUvr95Pi8OZ6qYIkTLJzNn3BGo8t2uBHsE7KKXmK6VWK6VWV1ZWJrEpIpO9u+kwP3llA49+uD3VTREiZZIZ7KuAYs/tYs/9AFrrBVrrSVrrSaWl7c72FSImxxpaATjeaE9xS4RInWQG+0XA+Z7bM4AlSXwtISJyOF0AZFlljIDIXB0O9kqpIUqp30ax6/PAAKXUBqAad/AXotM5XBoAm0VGGovMFXUhNK31cM/P3cDtbe3jud0CzI23gULEy+50B3vp2YtMJl0d0eUZaRybBHuRwSTYiy7vkQ+3AZLGEZlN3v2iy9PuLA52Tw9fiEwkwV50eYN75gMwrLRbilsiROpIsBdd3sCSPACsFsnZi8wlwV5kDI1OdROESBkJ9iJjuCRlLzKYBHvR5RkXaKVfLzKZBHvR5XmDvZZwLzKXBHvR5Rm5egn1IpNJsBeZQ6K9yGAS7EXGcEkaR2QwCfaiy5MLtEJIsBcZYOXuasAX9IU51Tbb+WRHyBpHIkEk2IuMIZOqUufWF9by7Ce729zn3je+5NpnVrJy1zEZOZUEEuxFl+bwK37mkviREnXNdt7ZeIgT7SwLuXjLUQCuXLCCv366JyGvve9YI3/4aId8eCDBXnRxzQ6/abPyB58S+6obcWkY2bewzf0cTt/v5743NyfktW/46yp+8+5Wjta1JOT50pkEe9GlNdud3tsS6lPjsUXbAWh2OCn/2dus3HUsZJ+jtc3UtzgCtt36wlrm/301e481xPzajS3u379TvtZJsBddW0Cwl7/3lHjvyyMA3PbSegBeWr3f+9j+6kY2Hajh9lc2hBz31oZDvL/5CIdrmmN+baPSqf+3hkwV9Rq0QqTazsp6FDC0A3Xp/f/IJW/b+WrC5OlfXXuAR644lWa7k7N/swTwlaEOp7bZEfGx9hiLkzU7nG3vmAGkZy/SxsyHlzLj4aUdOsY/vMs3+c5375tfAnDx+P4B2xtaHNz6wlrv/YrjTcwY2ZulPzmPJ685LWDfI7Wx9+wtyt2zb7FLyVMJ9qJL8581K7G+83liLf2KcwO2N7Y62XK4LmBbllUxuGcBc8b144lrJni3r9t/IubXt3oaID17Cfaii/PP3Egap/O5XJqBJXlkWQNXCTtU00TF8SZG9PGl5HZX+S7Ezh3Xn3X3zGbK0B5sO1KHK4qvZYdqmgLuV9a1UNPkTiP5X7vJVBLsRZcmAT61jta10LswhyxrYKh5+P1tAEwcXMJlp7pTPHnZgZcQu+dnU9Yjnw0VNVy1YEWbr/PxtkqmPriYtzYc9G47/X8+5FhDKwDNksaRYC+6tsCcvQT+zuRyaVbvPU7/7nlMKCsJeGzptkoA7rhoFNdPKwdg3IDikOeYOaoPAKv2VHPF05/xi9c38fTSndQ129lQccI7XHPtvuMAbDlUF/IcAC2SxpHROKJrC8jZp0Gsd7k09a0OCnNsKJXeC6T/5ZPdtDpcOF2ac0eUsuKOmby8ej8Pf7DNu09hjo0JZSU8Ne80zju5d8hzXDCmr/f2qt3VrPLUOXpw4RYAuudn8fdvT/b23HOzwvdfpWcvPXvRxQXk7FPXjKj9+t0tjLv3fRpao+uJtjpcHKltZvuR8D3aVDLGx599UikAfYtzuW7q4IB9jA+0C8f2IzfLGvZ5dj94Ea99dxrbfvU1XrjpDAb3zPc+dqLRzree/dybrmt1hA/qkrOXnr3o4tKpZ//upkM8/fEuwF1PpltO+3+eZ/56MZWeUgDb/+drIbnxVPrIk6q5YEwf77bu+dmMHVDEpgO1fPe8YVE9j1LKmwaaNqwXS38ynZpGO/k5Vk66ayHHGlrZ5bm429DqpKnVyYETjQHPIcFeevaiiwvs2Zs72n/nH75x53VRTiSq9Kv58tRHOxPepljZnS52HK0HINsWGGaMse99g4ZjdkRxfhZZVgtXTx4EwAeb3bN0/7x8Nz97dQOzHvk4YP+WCD3+TCLBXnRpgUMvU9eO9vhX5wSobWq7QiSEpiwe/mCbKS5E7q9u5J7XN3nvBwd744OsJD877tcaWJIfsu31dQdDtrVIz16CvejaAtM45o32S7ZWBtxviiI4feEZgeKvvTLCneHWF9byz1W++jfZQaklY0bs0NKCuF/rhmnl3OAZzdOWxxbviPu10l1UwV4plaWUerONx3OVUm8ppdYrpZ5TbgVKqdeVUp8opX6TuCYLET3/8G7iWM/CjYfIsVmYc0o/IPKFRoPWmu/984uQ7e0d1xm2Bl0sDh5VZNQrKsrNivu1CnJs3HjWkLifJ1bPr9zL0Dve5skl5v8waTfYK6XygDXA7DZ2mwdUaK3HAyWefa8FVmitzwTGKKVGJaC9QnRIOpRLmPPYMl794gAvzp/C92eeBLSfY952pN5bo/3+S8d4t6f6QqTD6Wp3mKPD5X48mgvQ0SgtzCE/28qA7qHF1O6eOzohrxHJXa9twqXhf9/b6t3mcumoZvx2tnaDvda6SWs9DqhoY7cZwAee24uB6cAJoJtSygrkAa1xtlWIDkuHnP2XB2sBWL//hDe/3V4P3X88+SWnDvDejib9k0w1UVxrKMpz9+gLEhTsc7OsbP7lhdw9N7Q/mWOz0L84F0sSpiwYdfqDDb3zHW5+fk3iXzBOicrZ9wRqPLdrgR7Aa8CFwE7gK611yFABpdR8pdRqpdTqysrK4IeFiJt/nt5pwmjvH9SrG+3keIJ9exdan1m2m2yrhWU/nU5xXhYv/NcZANTHUQ44Eeb9eVW7+7zynancM3d0yIXbeIUbdppjszBlWE/6FUcuoRyrR/wmhxleWePuExs1/M0kUf/bVYAx17nYc/8O4I9a63Kgh1JqWvBBWusFWutJWutJpaWlCWqKED7+36bNNiKj4ngjcx9f5r1f22T3BsC20ji3vrCW51bspdXpYlAP92iUAZ568Nc8s5KXPt+XxFZHduBEE18dcn9Led7z4RPO8N6FfDsJefZwHx49CrLJslgSvlLVgRNNIds+3VHlLQNhRokK9ouA8z23ZwBLgELAKETdAkS/4oQQCeLfs2+MclZqIv3qrc28/+XhgG3r959gT1UD9725mW1H3GPRR/YtZGhpgbdnHymN88A7X/HWhkMATB7Sw7u9tDDHe/vl1W1lXJNn+XZfoBvbv5hVd85k5Z0zO+31g0f9AEwoK8FqVd7rBIny2U7f0oqTBpdw+/kjsFoUTa2p/WbVlg4nzZRSQ4BbtNa3+21+HrhcKbUBWI87+O8AnldK3QLs82wTolP5d+gaOvkP8fM91TyzfDfPLN/NnofmeLf/9JUNbD1SF7A601+/NZm+xbne9E2knv0CzwzbUf2KAnrPuTZfqYFWZ2pG5Oyrds9aXXfPbIrzs4D4R9t0RJZfz/7sk3qxbHsVJflZZFkU9gQvS/iX5bu9t2+ZMZxDJ5q5Mqgyp9baVPWNog72Wuvhnp+7gduDHmsB5gYdsgc4M872CREX/1mzxuLTneXppb7LVLsq6xnSq4AnFu/wDk2sOO5LBfQocE8wMnqn4YK90Zs8ZUAxb9x6ZkAgsfhdgQw3KiXZLvjdx97zKkzAkMpY5PgF+wXXTaKqvgWlFDZr4tM404b1ZPOhWu+HePnP3g7Zp8XhiljvJxWkNo7o0vyvySb6q3x7/NdO3VfdSLbNElDx0Z+Rb1ZKkW2zhL1A+4+VewF4/OoJbfYYE33hsz1aa2+gv3pymXeR7842pJdvklZettV7PcNmUdgT/G2n1emiJL/tD7UWu7mCvcygFV2af7DvjKHPR+uaWb69CoBhfgujW5SKut5Nq8PF00t3BYzV3lVZz9sbDnHRKX0p79X2zNNFXx2NoeWxq6x3j/e/e+5oHrz8lE59bX/5nsVPgssc26yJD/Ytdle7H6pmKF3hT4K96NI6s1yC1prJ/7OIeX9eiculsQX1cDs64Wnone9427yz0l3V8ZtTy9s9rr7F0amTq3YedbctEeUP4vXZHTNY/v9mBGzrXZiLS+MtzJYIrU4XOba2e+1mq6EvwV50aYErVSX3tfyH3TXZnQE9u1aHy1vfPdj9l42N+JxHat295n+u2ofNohjZtzCqtmw+VMuOo3V8vqc6qv1jVdts5+o/uS9MnjaopJ29k69fcR69uuUEbBvRx/1/drQ2/P9/R72+7gAbKk4EXCMw/Pvmqd7f5/IdVQl5vUSRYC+6NKNnr1TylyX0T9Os3nucf/kNgWyyO7n5+bUhxwzqkcd1UwIX9HjILxVS3dDKHa9uYPGWo1w3dTDdo6wUefkfPmXWIx/zjac+42CYMeHBmlqdvLXhYIe/Eby70TestLidHHaqFOa60zu3v7w+7ud67rM9/ODFdeysbAiYpHdyn0J6dctm4uAenDfCPWfocE37/++dSYK96NKMNIhVqaSXS/B/+uv/EjiT9ERj+GohBdmhYyRmjPItz/f7D7d5K0ieOyK2iYef7TzGjqP1EdNYdqeLUfe8y60vfOEdwx+tiuPu4ZY/nj0iprZ1hjLPylYHI3yz6oi7X//Se3ui37q67/7wbFbdOQvAe2H4scU7TFVpVYK96NKMvzWrRSW9Zx9uhu68KWUAVAT1rm+b5Q6OV54+KOQY/2qQ73sW5Xj6uolh12iNxo9fXs+sR5ZGrMy4/Ygvl93YwbkIOyrrGdwzn+95CriZkf//ZzwFyhZ9FVgCwQjq4B5FZQkzCulQAj5gEkWCvejSXJ0Z7MOMjR83sDsAJxrcBcIG9chjz0Nz+MGsk9hy/4Vha7EHD9f70ewRAQtvR/LCf53Bk9ec5r3/tbF9A8bch8shH61t5ldvb/be78h/0c7Ket7ZeJiyHqELiJjVwx9sbX+nCG782+qA+8V5kdNWz95wOgCHE3SdIBEk2IsuzZvGsaikX6ANF+z7eZbeM8ah3/k1X2XG3CxrxPHyU4f29N6+KkzvP5xpw3sxZ1w/7rxoJL0Lc/jlpWMDPiQcTo3Wms92HuOlz/fx2KLtTH5gEZ/uPOYdRtiRkhJve1I+F4/vH/UxqZbIYalFeZGnKRnlK47WtkTcp7PJpCrRpfn37JOdPw1XcbJngfuPft3+E0Borz2S+y8by6xHlgLQu6hja7XOP2cY889xL+b9g1kn0a84l4+2HeWTHccYftfCkNmks0b14Y6LRjLz4aXsOFpPs90ZVTsbW51kWy1cMSm6DyMziHW8/WtfuC+233T2ED7fc5x1+0+02bPv7Qn2xhwEf1X1Ldz20jp+cfFo1u2vIcuqOHN4r5BRRIkmwV50cb4LtMnu2Yf7ym6UQTDkZEX3ZXp4b/eErKLc+P5Ei/OyuOmcoUwd1pO5jy/3Bvo3bj2TkX2LcLhc3slIAP9eW8Hlpw3gzOG9Qp6rttnO136/jDsvGsWccf1otjujPp9Uu2fuaH751mbGD+re4WMdThe3veQeyWN8GELbK231KMjGalHsqqxn6bZKhvfuhtOp+fnrm9h5tJ4DJ5oCFkW/5owyHviP5E5Ik2AvujQjwFs6oWd/uKaJsh753oJgAD27BQb7jizFt/6e87EkKJaOHVDMr79+CiX52eRmWb3XErL9Mrl9inI4UtvCrsr6sMF+w/4aDpxo4pYX1jJn3BxaHNF9AzCDb581hF++tZlX1x7gPycOZNqw0POLxL/ERUGOzZuuK2qjZ2+zWhhWWsCzn+zh2U/2MG1YTz71q5Rp6Fecy3+fM5QrokzVxUOCvejSHJ5on221JLVn73C6WLK1kvEDi9nnN48peEENo8cejUSPW7/y9LI2H58wqIR3vzzsna0bzH8VqiO1zTS1OslLk2Dv75o/reTL+y6IeqWsTQdqvLfvv3Qscx5zr0GQn932uVv8rscEB/oHLz+Fqye3/ftItPT4DiZEjJye4mfZNktSe/b1Le58/cntzHANN+vSLMYNcq8/tP1oXdjH/UtEH6ppptnuCqlDky46MiQy22phTP8i9jw0h5KCbG48eyhAuzn24Nc4d0Qpq38+i3/ceEanB3qQYC+6OIenjnmWNbk5+wbPKJZJg3vw3+cOjbifmeqbB/vuecO5fMIAdkfo2Te2+IL9ki1H2V3VkDZpHAgc1TTrkaU8uPCrdo/ZX93IkbrmgMXRbzxrCHsemtPuuWdZ3b9r47rLrFG96dUth7NOij6FlEgS7EWXZtSMz7JakjrO3giE+TlWzvLku40/9nTSuyiXgzXNrNx1LGTkSk2TL9g/umg7W4/UpVWwv2X68ID7Ty/dRfnP3ubDzZHXiz37N0vYdKCWkijLVPizeS642DypPP/VxFJBgr3osirrWnh00XbACPbJey1jfHpBts1bz92WqKurnWjaMPf4/isXrOD/vbLBu93udLHtSJ133oAh0YuCJFOeX479/dvO8da/N9bNDfae33KSsYziMVJc108tp39xLpPKe7RzRHKl37tRiCgd96tHk21NXs7+3U2HWLHLfQEuP9vq16NLv579OSNKKfSkLF794gBaa/ZUNXDmQ4t5e+OhkDLGa/YeT0UzY2IURPva2L6M6FPI4h+fS7bNQn2EEhH//dwa7+22xtRH8qdvTuKbUwfzvRnD+fSOmUkfR98eGY0jMoLNmpxyCQ0tDr7zD181y/xsm3cN2OCROLEEjFS4ZcZwHlq4BYAhd7wT8Ni8MwbzyY7QIYTpIMdmZc3PZ3mXTVRK0bswh80Hw/fsh5YWsKuygYJsK1OGdrxXflKfQn55aeTy1Z1NevYiI2RZLUmpevn44sDiYvk5Vu+iJcbPP17rrleTLjVk5p89NGwJhC/vu4CvndKPSzyPDe/dLaAWTzro2S0nYIWpcQOLqTje5B1NZWi2Ozl4oomrTh/El7+8kKGl0Q+ZNSvp2YuMkKwLtMGrHxVk27wjgIxXM0bg9Olg2YNUsVgUj189gcevngC4UzWf76n2jkt/7OoJPOZ5LN0V52Wzu6qBsb94j2U/ne6tZHnh7z+m2e6iX3HnL96eLNKzF6Z18EQT5T97myVbA4tXxZJ7z7Elp2fv/wFyw7RyivJs3hIJTZ6LtsaolhF90rN3OHFwCd85d1iqm5EUhX7lKC578hM2HaihsdXBnmPuWdBDTLDUYqJIsBemZRQP+9fn+wO2Rxu0/fdLVoljY4ZuYY6Ney8ZQ362jRLPzFcjFXLRKf24/9Ix/GCWeWu+Z6qT+/gmwR1raOXO1zZSWecuXtarWw7TT45twRgzkmAvTMuIzcHzkGIJ2skqcezw9NrrWhzeHrzNamH9L87n/kvHeF/7uqnl7S5QLTrfJacGXpsY0D2Pm/7urlv/+NUTvBdzuwIJ9sK0tCfrrVBB2zsuWWvQOvw+QfwXtyjOy/JOphHm5T9iKttmYeGmw2w7Us/4gcVMKk/9AuqJJO9GYVre2JyInr1SSZkA5P+c2Wk4rl7AfZeMYfboPtx1kW9hmT/MmxgydDbdyWgcYVqRQnMsHfTcLGvYlaTisWbv8YAx2hYT170RkV0/rZzrp5Wz95i7JtC4gcUByzl2FRLshWkZo26CQ2i0wd7/G0BettU7OiZRvv7HTwPuW8MsOC3Sx+CeBex5aE6qm5E0Xet7iuiSgitFRpvG8d+vrtlOk93JsTDLxCWKRYK9MDEJ9sK0jFgdHEOjDfb+u/1zlXv45gsr9yWiad5aOP4kjSPMTIK9MC3faJzg7dHx/1CYPboPAGU9E1Oy4KoFK0K2yfVZYWZRBXulVJZS6s02Hs9VSr2llFqvlHpOeb53K6V+qpRaoZRaqJTqeEFokdEideB1lNdZ/QffGLXMC+NcwLst0rMXZtZusFdK5QFrgNlt7DYPqNBajwdKgNlKqaHAGK31FGAhMDAB7RUZxDepKnicfcdz9kZRMqNuTSIZzetqQ/VE19Luu1Nr3aS1HgdUtLHbDOADz+3FwHRgJlCilPoYOBvYHWdbRYaJMMw+6pmwLr8djdryiRhrv7+6MeC+1RPts2zSsxfmlaiuSE/AWIK9FugBlAKVWutzcPfqzwo+SCk1Xym1Wim1urKyMkFNEV2Fd+hlzKNxfLeNgOxIQLAPXpDbWEQ8iaseChG3RAX7KqDYc7vYc78W2OrZtgsYEHyQ1nqB1nqS1npSaWnXKTgkEsNXIjhw+20vrYuqh+7/oWCMgU9EyYQWe+BFg56eFYgaWsKveCSEGSQq2C8CzvfcngEswZ3nn+TZNhx3wBciehHi8rLtVSGplHACc/but3oicvb2oA+aS0/tz9ShPfnhrBFxP7cQydLhYK+UGqKU+m3Q5ueBAUqpDUA1sEhr/RlwTCn1ObBVa70q/uaKTBJp6KX7sSiO90/jeHL2247URdg7ek5XYM++tDCHf86fQnmvrlP7XHQ9UY9D01oP9/zcDdwe9FgLMDfMMTfH20CRufxLHAcvWBLNAib+qR4jZ//0x7u4/YKT4xo5Yw/6dpArpYtFGpCxYsK0jJBqUaG16KPp2fsXPvOvWxNv3j74ekFutgR7YX4S7IVpxduzb7a7C5/deNYQ7zh7/+eNVfCInlyb/BkJ85N3qTAt/8lTwfE5moBt9Oyvn1ruzdlHe2xbnM7AnH1ulvTshflJsBem5QvKKiRARxOvjZ59bpbFm7OH+NM4Rs/eWL9Ugr1IBxLshWn5j7MPLpEQTcA2evY5WdaE5uyNYG+0KUsqoIk0IMFemJcnKFtUaOolmnht9OxzbJaAnH28k2iNC7TG80hNHJEOZKUqYVq+2jihPeeO5Oxzgi6gRnNxty3GxCzjG4IEe5EOJNgL0/KPyaE5+yjSOHYnOTZLmNo68bXL6XKhlK+HL2kckQ6kSyJMy1cILTS4R9uzD3fxNN6c/f7jTfTqlsM9c0czoHseA0q63uLUouuRnr0wLf8Sx7Hm7INTOBBfsNdas3RbJUN6FTBzVB9mjuoT83MJ0Zkk2AvT8l+8JDhAR5PGabY7w/bs4+nYv7KmguqGVqobWmN/EiFSQNI4wrQCh14Giibvfqimmd6FOSHb41nAZM3e4zEfK0QqSbAXpuU/aiY0jRMasN/ZeIg/frTTe7/F4WJImEqU8aRxmjzDOYVIN5LGEaanCO3ahwvXi7cc5dMdVdx83jAA/u+WM8N+KMSTxml1RLnauRAmIz17YVoBhdCiGI1jUaHpneBhlxBfz94I9i/cdEbMzyFEKkiwF6ZlBPjwM2hDA7ZFqagu3Da0ODlW3xJTm1ocLiYOLmHasF4xHS9EqkiwF6blPxonmr64ClP33t+sUb0BuOixZUz81Ye8s/FQh9sUaTinEGYn71phOk6X5l+f7w+oG789aDnBcDHdEqbuvb+TPFUqDR9vq+xw21ocLgn2Ii3JBVphOi+s3Mvdr39JnyL3sEkFXLlgRcA+rjBd+HArWvnzL4YGHR9Z43JpNh6o4fzRMpFKpB/pogjTqaxz59NrmuzuDWFKz0Tq2bd18TV47ViHM/y+Ty7ZQfnP3qb8Z2+z6UCNd3tVg7tdUr9epCMJ9sJ0Wj1BOMvifntawoyoCRfTlVJhe/wGR9AKU7YIBcz+972t3tsrdh3z3q6qc8+avXBs34ivIYRZSbAXpmP3BGUjGIcLyRFH47SRxrEHBXtrmA+RtlR5RvCUhpmVK4TZSbAXpuML9pHfnrGkcYaWdgu4n5vtS8cs317FtiN13m8Gl4zvD7iHaRqM9FKvbhLsRfqRYC9Mx8itGxdUw3XAwwV1i6XtC7TfnDo44L4xqqaqvoV5f17J+b/7mB+8tA6ACWXdybFZqG+x84vXN7HlcK23Z9+rW3aHz0mIVJPROMJ0QtM44WbBhh6n2unZB8+mNS7QXvfnVd5tb64/CEC/4jy65dh4eU0FJxrt/O2zvd59uuXIn41IP/KuFabjDfaeC7RR9+zbydn7Ky3M4WhdM+Bbq9Zf76IcjkUoYxyuBIMQZifBXpiOUYLYGBYf/QXa9uve/OfEgTS1Onl74yHe+/IIn+6oom9RLrurGgL2G9A9dPWpob0KGN2/KLqTEMJkJGcvTCckXIcrZham+KTFs8jJ4ZpmGlocYZ/7t98Yz5PXnua9v2xHFYdrmwP2mTOuH32KcgO29SzIjrgYihDpQIK9MB3f2rORh17+4MUvQra1Ol24NEx5cBHXPrMyqtfafLCW3VUNXHtGGf2Lc3ns6gk8cfUEAB696lTfcztcUipBpDV55wrTCc7EhEuRN7SG5tlfXl3hvb1u/4k2X+OJa9wBfamnPs70k3vz6R0zuWR8f++HzOQhPbz717U4ONbQKj17kbaiCvZKqSyl1JttPJ6rlHpLKbVeKfWc8ruCpZT6kVLqw0Q0VmQGI+9uvInCjcYJx2qJ/sLp3HH9GVrqW8WqpCArZB/jArE/GYkj0lW7wV4plQesAWa3sds8oEJrPR4oMfZVSg0Grk9AO0UGMYZVGl2GaGrUAxTmdiwQXzDGV/YgPzv02Owwk7rCLXMoRDpoN9hrrZu01uOAijZ2mwF84Lm9GJjuuf0ocEdcLRQZx0jjBP9sT47pbxGeAAAXyElEQVStYymWn5x/MqP7uUfXhEvPhKudU5QnPXuRnhL1zu0JGOUBa4GTlVLXAOuBzZEOUkrNB+YDlJWVJagpIv0FL0EYXbTvQBbHvb9F8fx/ncGbGw5S3jM/5PGsMD37vCwJ9iI9JeoCbRVQ7Lld7Lk/F5gJvAhMVErdGnyQ1nqB1nqS1npSaWlpgpoi0p2Rxtl+tB4IXwcnHGfQtNqVu45R02hv85iSgmy+ObU87ESpLE/Pfvyg7t5tedlygVakp0QF+0XA+Z7bM4AlWutrtNZnAVcBa7TWTyTotUQXFzwxKto0TnCwv3LBCr7YfzzmdiilWPbT6bz831O924IXQBEiXXQ42Culhiilfhu0+XlggFJqA1CNO/gLEZPg4N7erFiDM8x+2XGOix/UI59sm4WrJ7vTjMOCKmcKkS6iTkBqrYd7fu4Gbg96rAV32ibccXuAWbE3UWSa4JAdbRqnKDd0+GRHL9pG8uDlp/Dg5ack5LmESAWZVCVMJ/iCbFurTxmWba8MO5FKZrwK4SZ/CcJ0gtM2jiiC/fLtVWG3S7AXwk3+EoTpxJKzDzdMEuLP2QvRVchfgjCd4OD+7Cd72j0mN0uCvRBtkb8EYTrRDrX0F6lAWe/C3LDbhcg0EuyF6cQQ68NOiho7oKhDxdGE6Mok2AvTibY8gj9jxE6WXz0bZ5gFToTIVBLshelEMfgmhDNowROI7UNDiK5Kgr0wnViCtFEqwSqLgQsRlgR7YTrR9Oz7FOUEHhO0SLkQIpAEe2E6jnCriQcpLcxhf3UjD7zzFS6X9qZxLBb/NE7SmihE2pHi3MJ07I72o7TLBT/61zo+33Oci8f1D1tSIdoVroTIBNKzF6Zjj6Jn79Lam+5pdjhxau0eZukX36VnL4SPBHthOg5nFD17rb11b1rsLpwuuTgrRFsk2AvTsUcxQN6lfbNmNRqX1liC3s3SsRfCR3L2wnSiCfY7jtazw7Ns4YPvbGHzoVoAbDm+iC/j7IXwkZ69MB17hDTOj2aPCLvdCPTBJNQL4SM9e2E6kXr2pwwsDrsdYHDPfJpanTS0OJLVLCHSmvTshelEql8fabHvgmwrhbk2ThlQHNibl669EF4S7IXpRBp5GWm0TbbNgsOpsVhUwHBLifVC+EiwF6YTqWdvsSiunjwoZLvDqXG6dEjPXy7QCuEjwV6YjjNSsFcq7EQpu8uF06VDatdLqBfCR4K9MJ1IHXKrJXywd3pq49gsSkokCBGBBHuRNnIirCdrd2ocTo3VYgnM2UvcF8JLgr1IGzZr5J57i8MVcbSOEEKCvUgjtqCeu78WuxOrVQV8FEhKRwgfCfYibWQFBXN/zQ6nu2fvt0Nds0ywEsIgwV6kDZs19O2a7dlmd7pH4xgjckrys7jrolGd2j4hzEzKJYi0kWVRXHn6IF5ZU+HdlmOz0Oopr2CzKP598zTe2XiIH58/ImDxcSEynQR7kTaK87M4vahHwLacLAt1Le7bVouF0f2LGN2/KAWtE8LcJI0j0sK5I0rJsbnr1+dnW73bjW0QuXaOECLKYK+UylJKvdnG47lKqbeUUuuVUs8pz/dnpdTflFIrlFJvKKXkW4SImX9GJttvvL3/2PvgGbRCCJ92g71SKg9YA8xuY7d5QIXWejxQAsxWSp0F2LTWU4Ai4PwEtFd0ceEWDgfwD+NOv338A7/07IWIrN1gr7Vu0lqPAyra2G0G8IHn9mJgOnAEeDTa1xEC2q6LY/DfJaBnb5VgL0QkiUqt9ARqPLdrgZO11tsBlFL/AbiA94MPUkrNB+YDlJWVJagpIp05I/XsVfh9JGcvRHQS1eOuAoxlhIo991FKXQJ8H7hYax0yw0VrvUBrPUlrPam0tDRBTRHpLFJ5Y/9ETt/iXO/tnCz/nL18gRQikkT9dSzCl5OfASxRSvUFfgLM1VrXJeh1RBfniNCz9++0v3DTGRTnZQHQtyiXYaUFgPTshWhLh4O9UmqIUuq3QZufBwYopTYA1biD//VAP+A9pdRypdS3426t6PIiXqD1i+P9ivM4f3QfACaVl3D/pWMBqGmyJ719QqSrqHP2Wuvhnp+7gduDHmsB5gYd8mvPPyGiFjFnT2Cv3Zg1m2Oz8selOwFYtOUo3595UnIbKESakiSnMJWIo3GC3qktdiPYWyjMdfdZ6pqlZy9EJBLshalE27O/bEJ/AMYN6s5pZSUAfHPK4OQ2Tog0JrNahalECvZBsZ4Lx/Zjz0NzACjJzwZg+sjeyWyaEGlNevbCVFyu8NstbVSwNFI/be0jRKaTYC9MxREh2rcVxo0RPFIbR4jIJNgLU4k0qaqtTruR+ZFgL0RkEuyFqTjjSONIFkeIyCTYC1OJK40j0V6IiCTYC1OpbQq/SHhjqzPiMU7J2QvRLgn2wlSu/tOKsNvrWiJPmDLy/BYJ9kJEJMFemNKt04cH3K9rDt/jB79gL2kcISKSYC9MadzA4oD7bQX7LYfcRVUlZy9EZBLshSll2QLfmo2tkYP9q18cAELr5wghfOTPIw1UN7TS6ogwJrGLiraX7l8SWXr2QkQmtXFM7uZ/rGHhpsOcVtadV797Zqqb02mCFyKJtICV/0VZGY0jRGQS7E3iRGMrr31xgH7FuXznH2t5cf4UHv1wO5/tOgbA2n0nUtzC5NN+Ed0I3BblmyHbHiU9eyEikmBvEpc9+Ql7jjV671+1IHQIYrPdSW6WNWR7MjXbnXy09Sgj+xZR3qsgqa/lH9RtVnfgtlkstDpdtBXvf3/lqfz10z1JbZsQ6U5y9knQ0BL5YmKwvccamPHbjwICfbBzR7gXY39/85G429ZRd//fJr7zj7Wc99uPqG5oTepr+dfFGdKrGwDzPDXqI65DDlw2YQD/d0vmpLiEiIUE+wT7/YfbGPOL93jti4p29/3X5/s5938/YldVQ9jHB/fMZ89Dc/h/F44EYMeROq59ZkXEFZkaWx2MuvtdLn1iecS1XDtqy2HfWvEtjsizWBPBv5Z9j4Js9jw0h6smD0rqawqRKSTYJ9AX+47z+w+3A/DrhVsjBlytNYu3HOGn/97g3XZaWfeQ/aYM6QlA/+65ADy2eAef7DjGkq2VAfs5nC6O1jYz+p73aLI7WV9Rw98+24PTpQPy4LFotvsCfLInLYVrqu8VE/PhJUSmkpx9gnyw+Qg3/X014L64eLi2mbpmB8X5WSH7XrlgBat2V3vvP3rVqVx66gDKf/Y2AOMHdSfHZuHm84YBUJgb+Bzdcqw02538YckODpxo5t9rQ79F3PfmZu57czMAr99yJre9tI4zh/fiytMHMXZAccj+kTT79eYdCfq2EEm49WfloqsQiSHBPgEWbznCr97e7L0/bVhPlm2vojVMvd6Vu44FBPqn5k3kwrF9AfjLDZP4w5KdPHP9JLp7ltqD0CGFxxvsjLz73YBtw3t3Y/7ZQ/nGpIH8xx8+Zd1+3+idS5/8BIBdVQ08t2Ivq38+i17dcqI6t6ZW3zk4ItUfThAjZ/+TC072bpPRlEIkhgT7OGyoOMEPX1wXknMvLXQHUnuY4PjAwi0AXHNGGd89bxgDS/K9j80Y2YcZI/u0+7o/fnl9wP3LTu3PI1ec6h1zPqGsO+v2n+DZb53Ot579POT4ix9fzsCSPO6ZO4bR/YvaHJ/un8Z5+P1tPHb1hHbbFysj7ZXnN+LISB3FmY0SIuNJsO+AZruTZz/Zw4g+3XBpOFzbHBDobzxrCBPKutNid/Hq2gMBwd7udHHSXQsBd5Gv2/16r9G486KRPPDOloBtq+6aidOl6VGQHTC56KcXjOTaMwYzvHc33vvhOfzug23cd+kY7n3jSxZuOsyhmmYO1TRz8RPLAdhy/4URh3T6lyl4Y/3B5Ab7MCtOeYN90l5ViMwgF2g7YNn2Kn797hZu/Ntqbvr7anZV1gc8flpZCXPH9SfbU9fF7nTR6nDx6Ifbudpv3PxlEwZ0+LXnnzMs4P73Zgynd2Eu/YrzyLEFBuq8bCvDe7uHLp7ct5CnrptIn6JcHr5ivHcf/6qSI+9+l6O1zWEvKCc5TR/0Wkb1St82SdkLkRjSs49SXbOddzYeCtj27Cd7yLFZWHXXLBZuPMT5Y9wpGCNAffjVUbrnZfG7D7cB7vTOwh+cHXW+PNjqn8/itpfWsWx7FUNLOz7BKT/bxp++OQkFzBrdh1tnDPfm/ic/sAiAnQ9c5O1Zf7HveEztjEVjq4OLH3d/0/C/KGt8Y4l3VJEQmU6CfTtcLs3h2mYe+WAbr3mqKwYrzsviqsll3vvH6t2Tjx5a6Eu73Dp9OJdN6B9zoAfo1S2HS8b3Z9n2Kk4dVBLTc8we7bsmkJtl5enrJvKHj3ay3nNBd9id7wCw6s6Z3lIN/pZvr+Ksk3rF9Npt+dFL6zlU0xyy3ejlS6gXIj6SxmnH2xsPMe2hxbyyxj288cIxfbFZFPOmuIP7E9ecFnJMQ1A53qJcG7dfcDLDexfG3Z7/nDiQDfeez5AElS64YExfXr/lTP727ckB2yc/sIjfvLsVcF+LMMz788qEvG6wNX7fIppaO29svxCZQnr2bTha2xww8qUwx8Yf552GS7svIt550Sjys0P/C+dNGewNlADP3XhGwtqklKIoN3TsfrzGRRh736Mgm96FsX8biZZ/nr7er9yEEesliyNEfCTYt2HRlqMBdeRfvnkqSik8NbrCBnogIBh3ZEx7KnUPM/kL3L3s4MlOTpfG6dLeC9GJcKS2xXt7ZF/fNyDf0EuJ9kLEI6q/VqVUllLqzTYez1VKvaWUWq+Uek65hWxLXLOTr6q+hTte3Riw7eQ+0adhhnkuoCajF54MSikuOsU9uev7M0/ix7NHANBkd3LZqQMCxr7/8KV1zHpkaVImWRXm2LhgTF/vfUnjCJEY7QZ7pVQesAaY3cZu84AKrfV4oMSzb7htpqe15rT7P2DSrz4Meawjn1f/vGkKf75+UkJ7v8k2yDPBK8dmYeqwnt7t/bvn8eb3zvLef3P9QfZVN1LdGF0VzKO1zUz45fus23+i3Q+IvsW5AXMG5AKtEInRbhpHa90EjFNK7WhjtxnAvz23FwPTgcFhtr0fe1PDq2m0c+BEEwO651GUZ6PV6QoZd94Ra/edCCjl+9yNk8nPtgWU341G76JcZhblxtyOVDBq39gsit6F7rZnW90fVuEqbdY02r37GVwuzao91dQ02dlQcYI5p/Rnzd5qjjfaucxTtiF4EtdXh2q9t4P/lxXSsxciERKVs+8J1Hhu1wInR9iWcMt3VHHLC2sByM2y0Gx3MbAkj4rjTVx0Sl/mnzOM8QOL2+2VO5wuPt5eybf/ujpg+9knlSaj2aY0dkCR52cxA0ryuPy0AVx1unvU0W/f3xqy/7XPrOTjn04nN8uK1hq7U/P2xoPc9pLvovaTS3aGHHe8sZV+xXmAO132tUeXeR8Lyc1LrBciIRIV7KsAYzhHsed+tzDbAiil5gPzAcrKyoIfjsrEwSU8/I3xfHWolq1H6li2vYqK400AvLPxMO9sPMx9l4zh+mnlYY+vrGvhuRV7eWzRdu+2KycN4vO91ZwxpGfYY7qqy04dwPiB3Rla6p59+8gVp3ofe+jycZz9myUB+x+ta+HeN75k7b7jfOvMIQHXOJ6aN5FfvLEp4MKrYeqDi9n94EUopdhXHbhoy8xREWoDSR5HiLgkKtgvAs7HnbaZAfwOKAuzLYDWegGwAGDSpEkx/Tn3Lc7l6xMHBmxrbHWwu6qBDRU13PHqRn7xxpfsr27k8tMG0r97LtuP1nP3/20i22ZhQ0WN97gbppUzb0oZwzzBLs2uKcdNKeUN9MEG9cjnd1eO9/baR/Ur4qtDtbz4+X6AkIvZEweX8ODlp4R8UzK8sf4gs0f3YXelr7bQ/ZeN5ZrJgR/6SnL2QiREh4O9UmoIcIvW+na/zc8DlyulNgDrcQf/7DDbOkV+to0x/YsZ07+YgSV5XPfnVTyzfDfPLN8dsu+VkwYxblAx35g4KK0upqaC/6zdF/7rDF5esz+kONtts0ZQ32KnV7dsZozsw56H5rBmbzUuDd946jPvfj94cR1j+hfx5UF3vv7Ja05jzrh+Ia+ZWR+3QiRP1MFeaz3c83M3cHvQYy3A3KBDwm3rdGefVMqqu2ay7XA9J5pa2VfdyNBe3RjRpxs2i4WynvntP4kAYEivAn7z9XEU5tooKchm/jnDuHpyGduO1DH/72u45owyfjDrpJDjJg7uEfYCrxHo+xTlhA304F5wHOCkPuG/cQghoqPMMlll0qRJevXq8F/5hfkZ76O2Ul9bDtfy6toDfP20gTS2OjjRZKeu2cGsUb0jTlAD+GRHFWP6FwUs6CKEcFNKrdFaT2pvP5lBKxIimusbI/sWcedFRR1+7jOHJ77wmhCZRpLUQgiRASTYCyFEBpBgL4QQGUCCvRBCZAAJ9kIIkQEk2AshRAaQYC+EEBlAgr0QQmQA08ygVUrVAaF1dH2K8ZVMjkay9wfoRZhqngl8jVja1NFjkn0OsRwjv4vkPL8ZzyGW18jE91Nb+5+stW5/GT2ttSn+AavbeXxBB58vqftH0+YUtamjr5HUc5DfhZxDqs+jK7yf2to/2ramUxon4hq4Kdo/Fp3RpmSfR2e0SX4XyXl+M55DLK+Rie+nuM/BTGmc1TqKYj5mko5tDtYVzgG6xnnIOZhHOp1HtG01U89+QaobEIN0bHOwrnAO0DXOQ87BPNLpPKJqq2l69kIIIZLHTD17IYQQSSLBPgKl1N+UUiuUUm8opboppd5SSq1XSj2n3HLDbDtPKbXc82+/Uur6ND2PAqXU60qpT5RSv0nTcyhRSn3kOYe7U30O0P55ePbJUkq96XdMyLml7gxiO4dI21IljnPwPy7t1gKRYB+GUuoswKa1ngIUAd8GKrTW44ESYDYwL3ib1vojrfVZWuuzgA3AF6k5A7dYzwO4FlihtT4TGKOUGpWSEyCuc7gG+NJzDmcq99rJKRPNeSil8oA1uNtvCHduKRHrOUQ4r5SI4xyCjzu/0xsfJwn24R0BHvXctgD3Ah947i8GpgMzwmwDQCmVDwzXWm/ojMa2IdbzOAF0U0pZgTygtZPaG06s56CAQk9PTQGndlJ7I2n3PLTWTVrrcUCF33ER32cpENM5RDivVIn19xB8XNpJy0Ynm9Z6u9Z6lVLqPwAX7h66MXutFugB9AyzzTAbWNRJzY0ojvN4DbgQ2Al8pbXe2akN9xPHOfwD6A78G2jB/aGVMlGeRzhtvc86VRznYBqxnkOY497vlAYnkAT7CJRSlwDfBy4GDuOeroznZ5XnX/A2w8XAW53T0rbFeB53AH/UWpcDPZRS0zqzzcHi+F3cqLW+HHewP9ppDY4givMIp633WaeL8RxMJdZz8D9Oa+1IdjsTTYJ9GEqpvsBPgLla6zrcvXQjRzcDWBJhG560wXm4vxKmVBznUQg0e7a1AN06q83B4jiHc4CnlFI5uFM4Kzqz3cGiPI9wot0v6eI4B9OI9RzCHJd2JNiHdz3QD3hPKbUcyAIGKKU2ANW43yDPh9kGcDqwWWvdHPq0nS7W83gSuFkp9Rnu9EcqU1KxnsNCIBdYBtyvta5PReP9RHMe4UR6n6VCrOdgJrGeQ8BxSqlvd0prE0gmVQkhRAaQnr0QQmQACfZCCJEBJNgLIUQGkGAvhBAZQIK9yGhKqXuVUtuUUmuUUne1sd95SqnyzmuZEIklwV4I+CVwNnCFUmpOhH3OA8o7q0FCJJoEeyEArXUj8BJwnlJqoVLqM6XUswBKqb/jLpj1uFLqRc+2Pkqpd5VSK5VSd6Su5UJER4K9ED5VwETgj7gLjg1VSvXRWn8T+AvwPa31VZ597wBe1FqfAVyqlOqZkhYLESUJ9kL49AD2AtcBz+EupBapgNrJuGcZf4S7nET/zmigELFKuwL8QiSDp4b5FbhrAT0B/AtY6rdLE1Dg2VcBW4HXtdZLlFI3AMc7tcFCdJCUSxAZTSl1L+6FTmpxl0VeCzyFO6VjBX6itf5UKTUc+CvuWirXAA24Uzs9gB3ADelYCVFkDgn2QgiRASRnL4QQGUCCvRBCZAAJ9kIIkQEk2AshRAaQYC+EEBlAgr0QQmQACfZCCJEB/j9JJHVXVwFDGQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0xeed4390>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 将该策略的收益率转换为一个收益指数，并绘制一张图表\n",
    "to_index(trade_rets).plot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 230,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "D:\\Python\\anaconda\\lib\\site-packages\\ipykernel_launcher.py:3: FutureWarning: pd.rolling_std is deprecated for Series and will be removed in a future version, replace with \n",
      "\tSeries.rolling(window=250,min_periods=200,center=False).std()\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Date\n",
       "2012-07-23    0.231203\n",
       "2012-07-24    0.230438\n",
       "2012-07-25    0.230415\n",
       "2012-07-26    0.230871\n",
       "2012-07-27    0.231525\n",
       "Name: Adj Close, dtype: float64"
      ]
     },
     "execution_count": 230,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 假如你希望将该策略的性能按不同大小的交易期波幅进行划分。\n",
    "# 年度标准差是计算波幅的一种简单方法，可以同计算夏普比率来观察不同波动机制下的风险收益率\n",
    "vol = pd.rolling_std(returns,250,min_periods=200)*np.sqrt(250)\n",
    "vol.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 231,
   "metadata": {},
   "outputs": [],
   "source": [
    "def sharpe(rets,ann=250):\n",
    "    return rets.mean()/rets.std()*np.sqrt(ann)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 237,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "D:\\Python\\anaconda\\lib\\site-packages\\ipykernel_launcher.py:2: DeprecationWarning: \n",
      ".ix is deprecated. Please use\n",
      ".loc for label based indexing or\n",
      ".iloc for positional indexing\n",
      "\n",
      "See the documentation here:\n",
      "http://pandas.pydata.org/pandas-docs/stable/indexing.html#ix-indexer-is-deprecated\n",
      "  \n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Adj Close\n",
       "(0.0944, 0.16]   -0.012374\n",
       "(0.16, 0.188]     0.278558\n",
       "(0.188, 0.231]   -0.325217\n",
       "(0.231, 0.457]    0.525923\n",
       "Name: Adj Close, dtype: float64"
      ]
     },
     "execution_count": 237,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 利用qcut将vol划分为四等份，并用sharpe进行聚合\n",
    "trade_rets =trade_rets.ix[vol.index]  # 由于两个数据的长度不一样，将长度都设置为vol的长度\n",
    "trade_rets.groupby(pd.qcut(vol,4)).agg(sharpe)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 信号前沿分析\n",
    "介绍一种简化的截面动量投资组合，并得到模型参数化网格"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 将金融和技术领域中的几只股票做成一个投资组合，并加载它们的历史价格数据\n",
    "import pandas_datareader.data as web\n",
    "names = ['AAPL','GOOG','MSFT','GS','MS','BAC','C']\n",
    "def get_px(stock,start,end):\n",
    "    return web.get_data_yahoo(stock,start,end)['Adj Close']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": 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>AAPL</th>\n",
       "      <th>BAC</th>\n",
       "      <th>C</th>\n",
       "      <th>GOOG</th>\n",
       "      <th>GS</th>\n",
       "      <th>MS</th>\n",
       "      <th>MSFT</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</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>2008-12-31</th>\n",
       "      <td>10.555326</td>\n",
       "      <td>12.424986</td>\n",
       "      <td>60.725376</td>\n",
       "      <td>153.250580</td>\n",
       "      <td>71.672134</td>\n",
       "      <td>13.191438</td>\n",
       "      <td>14.868458</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2009-01-02</th>\n",
       "      <td>11.223147</td>\n",
       "      <td>12.645598</td>\n",
       "      <td>64.616867</td>\n",
       "      <td>160.060059</td>\n",
       "      <td>73.684967</td>\n",
       "      <td>13.972719</td>\n",
       "      <td>15.549164</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2009-01-05</th>\n",
       "      <td>11.696809</td>\n",
       "      <td>12.336741</td>\n",
       "      <td>64.073860</td>\n",
       "      <td>163.412491</td>\n",
       "      <td>75.400528</td>\n",
       "      <td>14.490844</td>\n",
       "      <td>15.694486</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2009-01-06</th>\n",
       "      <td>11.503882</td>\n",
       "      <td>12.601477</td>\n",
       "      <td>67.512863</td>\n",
       "      <td>166.406265</td>\n",
       "      <td>75.341080</td>\n",
       "      <td>16.102755</td>\n",
       "      <td>15.878043</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2009-01-07</th>\n",
       "      <td>11.255305</td>\n",
       "      <td>12.098475</td>\n",
       "      <td>64.707367</td>\n",
       "      <td>160.403763</td>\n",
       "      <td>71.765533</td>\n",
       "      <td>14.885596</td>\n",
       "      <td>14.921995</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 AAPL        BAC          C        GOOG         GS         MS  \\\n",
       "Date                                                                            \n",
       "2008-12-31  10.555326  12.424986  60.725376  153.250580  71.672134  13.191438   \n",
       "2009-01-02  11.223147  12.645598  64.616867  160.060059  73.684967  13.972719   \n",
       "2009-01-05  11.696809  12.336741  64.073860  163.412491  75.400528  14.490844   \n",
       "2009-01-06  11.503882  12.601477  67.512863  166.406265  75.341080  16.102755   \n",
       "2009-01-07  11.255305  12.098475  64.707367  160.403763  71.765533  14.885596   \n",
       "\n",
       "                 MSFT  \n",
       "Date                   \n",
       "2008-12-31  14.868458  \n",
       "2009-01-02  15.549164  \n",
       "2009-01-05  15.694486  \n",
       "2009-01-06  15.878043  \n",
       "2009-01-07  14.921995  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "px = pd.DataFrame({name:get_px(name,'2009-01-01','2012-06-01') for name in names})\n",
    "px.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "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>AAPL</th>\n",
       "      <th>BAC</th>\n",
       "      <th>C</th>\n",
       "      <th>GOOG</th>\n",
       "      <th>GS</th>\n",
       "      <th>MS</th>\n",
       "      <th>MSFT</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</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>2008-12-31</th>\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>2009-01-01</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2009-01-02</th>\n",
       "      <td>0.063269</td>\n",
       "      <td>0.017756</td>\n",
       "      <td>0.064083</td>\n",
       "      <td>0.044434</td>\n",
       "      <td>0.028084</td>\n",
       "      <td>0.059226</td>\n",
       "      <td>0.045782</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2009-01-05</th>\n",
       "      <td>0.042204</td>\n",
       "      <td>-0.024424</td>\n",
       "      <td>-0.008403</td>\n",
       "      <td>0.020945</td>\n",
       "      <td>0.023282</td>\n",
       "      <td>0.037081</td>\n",
       "      <td>0.009346</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2009-01-06</th>\n",
       "      <td>-0.016494</td>\n",
       "      <td>0.021459</td>\n",
       "      <td>0.053672</td>\n",
       "      <td>0.018320</td>\n",
       "      <td>-0.000788</td>\n",
       "      <td>0.111237</td>\n",
       "      <td>0.011696</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                AAPL       BAC         C      GOOG        GS        MS  \\\n",
       "Date                                                                     \n",
       "2008-12-31       NaN       NaN       NaN       NaN       NaN       NaN   \n",
       "2009-01-01  0.000000  0.000000  0.000000  0.000000  0.000000  0.000000   \n",
       "2009-01-02  0.063269  0.017756  0.064083  0.044434  0.028084  0.059226   \n",
       "2009-01-05  0.042204 -0.024424 -0.008403  0.020945  0.023282  0.037081   \n",
       "2009-01-06 -0.016494  0.021459  0.053672  0.018320 -0.000788  0.111237   \n",
       "\n",
       "                MSFT  \n",
       "Date                  \n",
       "2008-12-31       NaN  \n",
       "2009-01-01  0.000000  \n",
       "2009-01-02  0.045782  \n",
       "2009-01-05  0.009346  \n",
       "2009-01-06  0.011696  "
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# pad/ffill：用前一个非缺失值去填充该缺失值; \n",
    "# backfill/bfill：用下一个非缺失值填充该缺失值\n",
    "px = px.asfreq('B').fillna(method='pad')   # 工作日频率\n",
    "rets = px.pct_change()\n",
    "rets.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0xb1819d0>"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0xb1c0dd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "((1+rets).cumprod()-1).plot()   # 绘制每只股票的累计收益"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 对于投资组合的构建，需要计算特定回顾期的动量，然后按降序排列并标准化\n",
    "def calc_mom(price,lookback,lag):\n",
    "    mom_ret = price.shift(lag).pct_change(lookback)\n",
    "    ranks = mom_ret.rank(axis=1,ascending=False) # 每一行的数据按降序排序号，如最大值对于的序号为1，依次类推\n",
    "    # 标准化，公式（x-mean）/std\n",
    "    demeaned = ranks.sub(ranks.mean(axis=1), axis=0)\n",
    "    return demeaned.divide(demeaned.std(axis=1),axis=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 利用这个变换函数，再编写一个对策略进行事后检验的函数：通过指定回顾期和持有期（买卖之间的日数）计算投资组合整体的夏普比率\n",
    "compound = lambda x: (1+x).prod()-1\n",
    "daily_sr = lambda x: x.mean()/x.std()\n",
    "\n",
    "def strat_sr(prices,lb,hold):\n",
    "    # 计算投资组合权重\n",
    "    freq = '%dB' % hold    # 频率为hold个工作日\n",
    "    port = calc_mom(prices,lb,lag=1)\n",
    "    daily_rets = prices.pct_change()\n",
    "    \n",
    "    # 计算投资组合收益\n",
    "    port = port.shift(1).resample(freq,how='first')  # 按freq重采样，取第一个值\n",
    "    returns = daily_rets.resample(freq,how=compound)  # 按freq重采样后，最终值由compound方法计算得到\n",
    "    port_rets = (port*returns).sum(axis=1)  # 两个数据相同日期上的值相乘完后，按行求和\n",
    "    \n",
    "    return daily_sr(port_rets)*np.sqrt(252/hold)  # 输出的是一个数值\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "D:\\Python\\anaconda\\lib\\site-packages\\ipykernel_launcher.py:12: FutureWarning: how in .resample() is deprecated\n",
      "the new syntax is .resample(...).first()\n",
      "  if sys.path[0] == '':\n",
      "D:\\Python\\anaconda\\lib\\site-packages\\ipykernel_launcher.py:13: FutureWarning: how in .resample() is deprecated\n",
      "the new syntax is .resample(...)..apply(<func>)\n",
      "  del sys.path[0]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.13145038363829603"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 通过价格数据以及一对参数组合调用该函数将会得到一个标量值\n",
    "strat_sr(px,70,30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "D:\\Python\\anaconda\\lib\\site-packages\\ipykernel_launcher.py:12: FutureWarning: how in .resample() is deprecated\n",
      "the new syntax is .resample(...).first()\n",
      "  if sys.path[0] == '':\n",
      "D:\\Python\\anaconda\\lib\\site-packages\\ipykernel_launcher.py:13: FutureWarning: how in .resample() is deprecated\n",
      "the new syntax is .resample(...)..apply(<func>)\n",
      "  del sys.path[0]\n"
     ]
    }
   ],
   "source": [
    "# 然后对参数网格（即多对参数组合）应用strat_sr函数，并将结果保存在一个defaultdict中，最后再将全部结果放进一个DataFrame\n",
    "from collections import defaultdict\n",
    "\n",
    "lookbacks = range(20,90,5)\n",
    "holdings = range(20,90,5)\n",
    "dd = defaultdict(dict)\n",
    "for lb in lookbacks:\n",
    "    for hold in holdings:\n",
    "        dd[lb][hold] = strat_sr(px,lb,hold)  # dd[lb] 输出是：{lb:{} }  相当于lb是列，hold是对于的行\n",
    "ddf = pd.DataFrame(dd)\n",
    "ddf.index.name = 'Holding Period'\n",
    "ddf.columns.name = 'Lookback Period'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 为了便于观察，将该结果图形化，生成一张带有装饰物的热图\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "def heatmap(df,cmap=plt.cm.gray_r):\n",
    "    fig = plt.figure()\n",
    "    ax = fig.add_subplot(111)\n",
    "    axim = ax.imshow(df.values,cmap = cmap, interpolation='nearest')\n",
    "    ax.set_xlabel(df.columns.name)\n",
    "    ax.set_xticks(np.arange(len(df.columns)))\n",
    "    ax.set_xticklabels(list(df.columns))\n",
    "    ax.set_ylabel(df.index.name)\n",
    "    ax.set_yticks(np.arange(len(df.index)))\n",
    "    ax.set_yticklabels(list(df.index))\n",
    "    plt.colorbar(axim)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0xafd01d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 对事后检验结果调用该函数\n",
    "heatmap(ddf)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 期货合约转仓\n",
    "期货是一种无所不在的衍生品合约。它是一种在指定日期交收指定资产（比如石油、黄金或FTSE100指数的股份）的约定。在实践中，由于期货合约具有限时性，对（股票、货币、商品、债券以及其他资产类）期货合约的建模和交易是复杂的。例如，对于某种期货（比如银或铜期货），在给定时间点，可能有多个到期时间不同的合约被交易。\n",
    "- 一般来说，下一个期满的期货合约（即近期合约）将是最具流动性的（成交量最高和买卖差价最低）\n",
    "- 通过一个表示盈亏（始终持有近期合约）的连续的收益指数即可轻松实现建模和预测\n",
    "- 从一份到期合约过渡到下一期（或更远）合约称为转仓\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Date\n",
       "2011-08-01    1074.033051\n",
       "2011-08-02    1046.594086\n",
       "2011-08-03    1052.265320\n",
       "2011-08-04    1002.975693\n",
       "2011-08-05    1001.474686\n",
       "2011-08-08     936.255035\n",
       "2011-08-09     979.790573\n",
       "2011-08-10     936.505127\n",
       "2011-08-11     978.539124\n",
       "2011-08-12     985.127945\n",
       "2011-08-15    1005.977859\n",
       "2011-08-16     997.387695\n",
       "2011-08-17     998.054886\n",
       "2011-08-18     955.020218\n",
       "2011-08-19     939.424057\n",
       "2011-08-22     940.175095\n",
       "2011-08-23     971.116486\n",
       "2011-08-24     984.794006\n",
       "2011-08-25     969.782181\n",
       "2011-08-26     983.876724\n",
       "2011-08-29    1012.149658\n",
       "2011-08-30    1014.818573\n",
       "2011-08-31    1019.322205\n",
       "2011-09-01    1008.646774\n",
       "2011-09-02     982.875977\n",
       "2011-09-06     975.703506\n",
       "2011-09-07    1003.225632\n",
       "2011-09-08     992.800674\n",
       "2011-09-09     966.779556\n",
       "2011-09-12     973.034668\n",
       "                 ...     \n",
       "2012-06-15    1142.504120\n",
       "2012-06-18    1144.718475\n",
       "2012-06-19    1155.790939\n",
       "2012-06-20    1153.917084\n",
       "2012-06-21    1128.024673\n",
       "2012-06-22    1136.711884\n",
       "2012-06-25    1118.485641\n",
       "2012-06-26    1124.106598\n",
       "2012-06-27    1134.242249\n",
       "2012-06-28    1131.005783\n",
       "2012-06-29    1159.197998\n",
       "2012-07-02    1162.690125\n",
       "2012-07-03    1170.355453\n",
       "2012-07-05    1165.074615\n",
       "2012-07-06    1154.002686\n",
       "2012-07-09    1152.554398\n",
       "2012-07-10    1142.504120\n",
       "2012-07-11    1142.674179\n",
       "2012-07-12    1137.138214\n",
       "2012-07-13    1156.216736\n",
       "2012-07-16    1153.491287\n",
       "2012-07-17    1161.412430\n",
       "2012-07-18    1170.014877\n",
       "2012-07-19    1173.081360\n",
       "2012-07-20    1162.349243\n",
       "2012-07-23    1150.595779\n",
       "2012-07-24    1140.715332\n",
       "2012-07-25    1140.971146\n",
       "2012-07-26    1159.793777\n",
       "2012-07-27    1181.172333\n",
       "Name: Adj Close, Length: 251, dtype: float64"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 用SPY交易所基金的部分价格作为标准普尔500 指数的代理\n",
    "import pandas_datareader.data as web\n",
    "# 标准普尔500指数的近似价格\n",
    "px = web.get_data_yahoo('SPY','2011-08-01','2012-07-27')['Adj Close']*10\n",
    "px"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "ESU2   2012-09-21\n",
       "ESZ2   2012-12-21\n",
       "dtype: datetime64[ns]"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 在一个Series中放了两份标准普尔500指数期货合约及其到期日期\n",
    "from datetime import datetime\n",
    "expiry = {'ESU2':datetime(2012,9,21),\n",
    "          'ESZ2':datetime(2012,12,21)}\n",
    "expiry = pd.Series(expiry).sort_values()  # series的数据 按值升序排序\n",
    "expiry"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "323.5 -38.0\n"
     ]
    }
   ],
   "source": [
    "# 用Yahoo!Finance的价格以及一个随机漫步和一些噪声来模拟这两份合约未来的走势\n",
    "np.random.seed(12347)\n",
    "N = 200\n",
    "walk = (np.random.randint(0,200,size=N)-100)*0.25   # 数据范围在-25 到 25之间\n",
    "perturb = (np.random.randint(0,20,size=N)-10)*0.25  # 数据范围在-2.5 到 2.5之间\n",
    "walk = walk.cumsum()  # 累计求和"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "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>ESU2</th>\n",
       "      <th>ESZ2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2011-08-01</th>\n",
       "      <td>1074.033051</td>\n",
       "      <td>1074.033051</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2011-08-02</th>\n",
       "      <td>1046.594086</td>\n",
       "      <td>1046.594086</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2011-08-03</th>\n",
       "      <td>1052.265320</td>\n",
       "      <td>1052.265320</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2011-08-04</th>\n",
       "      <td>1002.975693</td>\n",
       "      <td>1002.975693</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2011-08-05</th>\n",
       "      <td>1001.474686</td>\n",
       "      <td>1001.474686</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                   ESU2         ESZ2\n",
       "2011-08-01  1074.033051  1074.033051\n",
       "2011-08-02  1046.594086  1046.594086\n",
       "2011-08-03  1052.265320  1052.265320\n",
       "2011-08-04  1002.975693  1002.975693\n",
       "2011-08-05  1001.474686  1001.474686"
      ]
     },
     "execution_count": 95,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rng = pd.date_range(px.index[0],periods = len(px)+N,freq='B')\n",
    "near = np.concatenate([px.values,px.values[-1]+walk])  # 近期合约\n",
    "far = np.concatenate([px.values,px.values[-1]+walk+perturb])  # 远期合约\n",
    "prices = pd.DataFrame({'ESU2':near,'ESZ2':far},index=rng)  # prices是关于两个合约的时间序列\n",
    "prices.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 将多个时间序列合并为单个连续序列的一个方法就是构造一个加权矩阵。活动合约的权重应该设为1，直到满期为值。到那个时候，必须决定一个转仓约定\n",
    "# 下面这个函数可以计算一个加权矩阵（权重根据到期前的期数减少而线性衰减）\n",
    "def get_roll_weights(start,expiry,items,roll_periods=5):\n",
    "    #start:用于计算权重加权矩阵的第一天\n",
    "    #expiry：由“合约代码->到期日期”组成的序列\n",
    "    # items: 一组合约名称\n",
    "    \n",
    "    dates = pd.date_range(start,expiry[-1],freq='B')\n",
    "    weights = pd.DataFrame(np.zeros((len(dates),len(items))),index=dates,columns=items)\n",
    "    \n",
    "    prev_date = weights.index[0]\n",
    "    for i,(item,ex_date) in enumerate(expiry.iteritems()):\n",
    "        if i < len(expiry) - 1:  # 不是最后一个合约\n",
    "             # prev_date到 到期时间前的所有数据权重设为1\n",
    "            weights.ix[prev_date:pd.Timestamp(ex_date)-pd.offsets.BDay(),item] = 1 \n",
    "            roll_rng = pd.date_range(end=pd.Timestamp(ex_date)-pd.offsets.BDay(),periods=roll_periods+1,freq='B')\n",
    "            decay_weights = np.linspace(0,1,roll_periods+1)\n",
    "            weights.ix[roll_rng,item] = 1-decay_weights  # 当前合约在roll_rng时间上随着时间减少权重\n",
    "            weights.ix[roll_rng,expiry.index[i+1]] = decay_weights # 下个合约在roll_rng时间上随着时间增加权重\n",
    "        else:  # 是最后一个合约\n",
    "            weights.ix[prev_date:,item] = 1\n",
    "        prev_date = ex_date\n",
    "    return weights"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "D:\\Python\\anaconda\\lib\\site-packages\\ipykernel_launcher.py:15: DeprecationWarning: \n",
      ".ix is deprecated. Please use\n",
      ".loc for label based indexing or\n",
      ".iloc for positional indexing\n",
      "\n",
      "See the documentation here:\n",
      "http://pandas.pydata.org/pandas-docs/stable/indexing.html#ix-indexer-is-deprecated\n",
      "  from ipykernel import kernelapp as app\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>ESU2</th>\n",
       "      <th>ESZ2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2012-09-12</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-09-13</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-09-14</th>\n",
       "      <td>0.8</td>\n",
       "      <td>0.2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-09-17</th>\n",
       "      <td>0.6</td>\n",
       "      <td>0.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-09-18</th>\n",
       "      <td>0.4</td>\n",
       "      <td>0.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-09-19</th>\n",
       "      <td>0.2</td>\n",
       "      <td>0.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-09-20</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-09-21</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            ESU2  ESZ2\n",
       "2012-09-12   1.0   0.0\n",
       "2012-09-13   1.0   0.0\n",
       "2012-09-14   0.8   0.2\n",
       "2012-09-17   0.6   0.4\n",
       "2012-09-18   0.4   0.6\n",
       "2012-09-19   0.2   0.8\n",
       "2012-09-20   0.0   1.0\n",
       "2012-09-21   0.0   1.0"
      ]
     },
     "execution_count": 136,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 快到ESU2到期日的那几天的权重\n",
    "weights = get_roll_weights('6/1/2012',expiry,prices.columns)\n",
    "weights.ix['2012-09-12':'2012-09-21']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 139,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2012-09-03    0.009396\n",
       "2012-09-04    0.002234\n",
       "2012-09-05   -0.007616\n",
       "2012-09-06    0.002059\n",
       "2012-09-07    0.013449\n",
       "2012-09-10   -0.005161\n",
       "2012-09-11   -0.001667\n",
       "2012-09-12    0.003526\n",
       "2012-09-13   -0.011835\n",
       "2012-09-14    0.002208\n",
       "2012-09-17    0.010161\n",
       "2012-09-18   -0.000813\n",
       "2012-09-19   -0.004479\n",
       "2012-09-20    0.005951\n",
       "2012-09-21    0.008503\n",
       "2012-09-24    0.015947\n",
       "2012-09-25    0.016959\n",
       "2012-09-26    0.014902\n",
       "2012-09-27    0.016956\n",
       "2012-09-28    0.003094\n",
       "Freq: B, dtype: float64"
      ]
     },
     "execution_count": 139,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 转仓期货收益就是合约收益的加权和\n",
    "rolled_returns = (prices.pct_change()*weights).sum(1)\n",
    "rolled_returns['2012-09']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 移动相关系数与线性回归\n",
    "移动窗口和指数加权时间序列函数就是用于处理动态模型的工具。\n",
    "- 相关系数是观察两个资产时间序列的变化的协动性的一种手段。pandas的rolling_corr函数可以根据两个收益序列计算出移动窗口相关系数"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 258,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 从Yahoo!Finance加载一些价格序列，并计算每日收益率\n",
    "aapl = web.get_data_yahoo('AAPL','2000-01-01','2012-07-27')['Adj Close']\n",
    "msft = web.get_data_yahoo('MSFT','2000-01-01','2012-07-27')['Adj Close']\n",
    "aapl_rets = aapl.pct_change()\n",
    "msft_rets = msft.pct_change()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 259,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "D:\\Python\\anaconda\\lib\\site-packages\\ipykernel_launcher.py:2: FutureWarning: pd.rolling_corr is deprecated for Series and will be removed in a future version, replace with \n",
      "\tSeries.rolling(window=250).corr(other=<Series>)\n",
      "  \n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x13a2b050>"
      ]
     },
     "execution_count": 259,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0xe69ef90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 计算一年期移动相关系数并绘制图标\n",
    "pd.rolling_corr(aapl_rets,msft_rets,250).plot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 260,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 两个资产之间的相关系数存在一个问题，即不能捕获波动性差异。\n",
    "# 最小二乘回归提供了另一种对一个变量与一个或多个其他预测变量之间的动态关系的建模办法\n",
    "msft_rets.dropna(inplace=True)\n",
    "aapl_rets.dropna(inplace=True)\n",
    "x= pd.DataFrame({'MSFT':msft_rets})\n",
    "\n",
    "import statsmodels.api as sm\n",
    "res=[]\n",
    "timeIndex = []\n",
    "for i in range(len(x)):\n",
    "    if i >=250:\n",
    "        windows_x =  x[i-250:i]\n",
    "        windows_X = sm.add_constant(windows_x)  # 增加常数项\n",
    "        windows_y =  aapl_rets[i-250:i]\n",
    "        res_model = sm.OLS(windows_y,windows_X).fit().params\n",
    "        res.append([res_model['const'],res_model['MSFT']])\n",
    "        timeIndex.append(x.index[i])\n",
    "results  = pd.DataFrame(res,columns=['const','MSFT'],index = timeIndex)  \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 261,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x13a26750>"
      ]
     },
     "execution_count": 261,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x13a2bfb0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "results['MSFT'].plot()"
   ]
  }
 ],
 "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": "3.6.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}