Assignment3_Kyrkos_Konstantinos.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Assignment III: Clustering and Classification of MEP's Tweets\n",
    "\n",
    "---\n",
    "\n",
    ">Konstantinos Kyrkos <br>\n",
    ">Registration Number p3351907 <br>\n",
    ">kkyrkos@aueb.gr"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This notebook provides my solutions for the third assignment for the Practical Data Science class at Data Science MSc at AUEB. <br>\n",
    "Bollow you will find the solution to all 3 parts of the assignment. <br>\n",
    "The only prerequisites for running the commands below are providing the retweets.csv file found in the following link and a config file named \"twitter_config.py\" which is used in order to connect and use the Twitter Api.\n",
    "https://www.clarin.si/repository/xmlui/handle/11356/1071#\n",
    "\n",
    "## Data Preparation\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We start by importing the libraries we will use in the following exercises."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import tweepy\n",
    "import json\n",
    "from twitter_config import config\n",
    "from yellowbrick.cluster import KElbowVisualizer\n",
    "from sklearn.cluster import KMeans\n",
    "from sklearn import metrics\n",
    "from yellowbrick.cluster import SilhouetteVisualizer\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn import preprocessing\n",
    "from sklearn.linear_model import SGDClassifier\n",
    "from sklearn import metrics\n",
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.metrics import classification_report, silhouette_score, confusion_matrix, accuracy_score\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.tree import DecisionTreeRegressor\n",
    "from sklearn.model_selection import KFold\n",
    "from scipy.stats import sem\n",
    "from sklearn.ensemble import BaggingClassifier\n",
    "from sklearn.ensemble import ExtraTreesClassifier\n",
    "from scipy.stats import sem\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.naive_bayes import GaussianNB\n",
    "from sklearn.dummy import DummyClassifier\n",
    "from sklearn.naive_bayes import MultinomialNB\n",
    "from yellowbrick.cluster import InterclusterDistance\n",
    "from sklearn.feature_extraction.text import TfidfTransformer\n",
    "import matplotlib.pyplot as plt\n",
    "import xgboost as xgb"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next, we import the file we will use to extract the retweets and then we filter the data imported to keep only the tweets written in English."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = pd.read_csv(\"retweets.csv\") \n",
    "data.head()\n",
    "\n",
    "data = data.loc[data['lang'] == 'en']\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We will use describe method to have an overview of the data we have to manage."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "         origUserId      origMepId  origMepGroupId  origMepCountryId  \\\n",
      "count  1.379600e+04   13796.000000    13796.000000      13796.000000   \n",
      "mean   4.823502e+08   73536.347202        3.516527         20.398884   \n",
      "std    7.144207e+08   50570.469713        1.997484          8.856113   \n",
      "min    9.234892e+06     840.000000        0.000000          0.000000   \n",
      "25%    1.976367e+07    4555.000000        1.000000         13.000000   \n",
      "50%    1.444836e+08   96907.000000        4.000000         27.000000   \n",
      "75%    8.198434e+08  124940.000000        6.000000         27.000000   \n",
      "max    3.520822e+09  130833.000000        8.000000         27.000000   \n",
      "\n",
      "       retweetUserId   retweetMepId  retweetMepGroupId  retweetMepCountryId  \\\n",
      "count   1.379600e+04   13796.000000       13796.000000          13796.00000   \n",
      "mean    6.329301e+08   99903.399681           3.536750             20.63185   \n",
      "std     8.118056e+08   40182.608699           2.035608              8.60228   \n",
      "min     9.234892e+06     840.000000           0.000000              0.00000   \n",
      "25%     1.061105e+08   96805.000000           1.000000             14.00000   \n",
      "50%     2.430972e+08  124896.000000           4.000000             26.00000   \n",
      "75%     8.290609e+08  124960.000000           6.000000             27.00000   \n",
      "max     3.889367e+09  134243.000000           8.000000             27.00000   \n",
      "\n",
      "        origTweetId  retweetTweetId  \n",
      "count  1.379600e+04    1.379600e+04  \n",
      "mean   6.070628e+17    6.073327e+17  \n",
      "std    4.869547e+16    4.848900e+16  \n",
      "min    2.204733e+17    5.176455e+17  \n",
      "25%    5.674471e+17    5.678041e+17  \n",
      "50%    6.082114e+17    6.082721e+17  \n",
      "75%    6.514254e+17    6.516432e+17  \n",
      "max    6.913697e+17    6.915363e+17  \n"
     ]
    }
   ],
   "source": [
    "print(data.describe())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next we set the parameters needed in order to access the tweeter api."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "auth = tweepy.OAuthHandler(config['consumer_key'], config['consumer_secret'])\n",
    "auth.set_access_token(config['access_token'], config['access_token_secret'])\n",
    "\n",
    "api = tweepy.API(auth)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "At this point we can access the tweeter api using the call statuses_lookup and feeding it with 100 tweet ids per time. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "tweet_set = []\n",
    "entries = len(data.origTweetId)\n",
    "tweet_IDs = data.origTweetId\n",
    "\n",
    "try:\n",
    "    for i in range((entries // 100) + 1):\n",
    "        end_loc = min((i + 1) * 100, entries)\n",
    "        tweet_cds = tweet_IDs.iloc[i * 100:end_loc].tolist()\n",
    "        tweets = api.statuses_lookup(id_=tweet_cds)\n",
    "        tweet_set.append(tweets)\n",
    "    \n",
    "except tweepy.TweepError:\n",
    "    print ('Something went wrong, quitting...')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next, we want to remove the empty tweets from the ones we got and store them in a more usefull manner. After that we print one tweet to see its format."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'created_at': 'Mon Oct 06 09:05:14 +0000 2014', 'id': 519050739642286080, 'id_str': '519050739642286080', 'text': \"Lord Hill's written answers to EP are now on my blog. No transparency on his lobbying past. Shame! http://t.co/OT7TGKGDek #EPhearings2014\", 'truncated': False, 'entities': {'hashtags': [{'text': 'EPhearings2014', 'indices': [122, 137]}], 'symbols': [], 'user_mentions': [], 'urls': [{'url': 'http://t.co/OT7TGKGDek', 'expanded_url': 'http://www.sven-giegold.de/2014/antworten-der-kommissionskandidatinnen', 'display_url': 'sven-giegold.de/2014/antworten…', 'indices': [99, 121]}]}, 'source': '<a href=\"http://twitter.com\" rel=\"nofollow\">Twitter Web Client</a>', 'in_reply_to_status_id': None, 'in_reply_to_status_id_str': None, 'in_reply_to_user_id': None, 'in_reply_to_user_id_str': None, 'in_reply_to_screen_name': None, 'user': {'id': 16705726, 'id_str': '16705726', 'name': 'Sven Giegold', 'screen_name': 'sven_giegold', 'location': 'Düsseldorf', 'description': 'Spitzenkandidat von Bündnis 90/Die Grünen für die Europawahl. Demokratie, EU-Wirtschafts- & Finanzmarktpolitik, Solidarische Ökonomie, Green New Deal, NRW-LaVo.', 'url': 'https://t.co/BlxwiKBPnd', 'entities': {'url': {'urls': [{'url': 'https://t.co/BlxwiKBPnd', 'expanded_url': 'http://www.sven-giegold.de', 'display_url': 'sven-giegold.de', 'indices': [0, 23]}]}, 'description': {'urls': []}}, 'protected': False, 'followers_count': 40232, 'friends_count': 782, 'listed_count': 1144, 'created_at': 'Sun Oct 12 07:22:18 +0000 2008', 'favourites_count': 2271, 'utc_offset': None, 'time_zone': None, 'geo_enabled': True, 'verified': True, 'statuses_count': 11654, 'lang': None, 'contributors_enabled': False, 'is_translator': False, 'is_translation_enabled': False, 'profile_background_color': '352726', 'profile_background_image_url': 'http://abs.twimg.com/images/themes/theme5/bg.gif', 'profile_background_image_url_https': 'https://abs.twimg.com/images/themes/theme5/bg.gif', 'profile_background_tile': False, 'profile_image_url': 'http://pbs.twimg.com/profile_images/1604429416/sg_ism001_kl_normal.jpg', 'profile_image_url_https': 'https://pbs.twimg.com/profile_images/1604429416/sg_ism001_kl_normal.jpg', 'profile_banner_url': 'https://pbs.twimg.com/profile_banners/16705726/1558937393', 'profile_link_color': 'D02B55', 'profile_sidebar_border_color': '829D5E', 'profile_sidebar_fill_color': '99CC33', 'profile_text_color': '3E4415', 'profile_use_background_image': True, 'has_extended_profile': False, 'default_profile': False, 'default_profile_image': False, 'can_media_tag': True, 'followed_by': False, 'following': False, 'follow_request_sent': False, 'notifications': False, 'translator_type': 'none'}, 'geo': None, 'coordinates': None, 'place': None, 'contributors': None, 'is_quote_status': False, 'retweet_count': 32, 'favorite_count': 6, 'favorited': False, 'retweeted': False, 'possibly_sensitive': False, 'lang': 'en'}\n"
     ]
    }
   ],
   "source": [
    "n_tweets = tweet_set\n",
    "filtered_tweets = []\n",
    "\n",
    "for tweet_group in n_tweets: \n",
    "    for tweet in tweet_group:\n",
    "        if not (tweet._json['text'].isspace()):\n",
    "            filtered_tweets.append(tweet._json)\n",
    "        \n",
    "print(filtered_tweets[0])\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we will put the tweets we found to a new column in a new copy of our original dataframe in order to keep the original for any later use. Then we will print the first 5 rows of the new dataframe in order to confirm that the new column has been added correctly."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "    origUserId origUserScreenName  origMepId    origMepName  origMepGroupId  \\\n",
      "3     19017675       Nigel_Farage       4525   Nigel FARAGE               6   \n",
      "5     17675072       MartinSchulz       1911  Martin SCHULZ               1   \n",
      "11    17675072       MartinSchulz       1911  Martin SCHULZ               1   \n",
      "12   744743599           Tim_Aker      99650       Tim AKER               6   \n",
      "13   744743599           Tim_Aker      99650       Tim AKER               6   \n",
      "\n",
      "   origMepGroupShort  origMepCountryId origMepCountryShort  retweetUserId  \\\n",
      "3               EFDD                27                 GBR      121171051   \n",
      "5                S&D                10                 DEU     1668992125   \n",
      "11               S&D                10                 DEU       21648649   \n",
      "12              EFDD                27                 GBR      121171051   \n",
      "13              EFDD                27                 GBR       19017675   \n",
      "\n",
      "   retweetUserScreenName  ...  retweetMepGroupId retweetMepGroupShort  \\\n",
      "3         MargotLJParker  ...                  6                 EFDD   \n",
      "5          NathanGillMEP  ...                  6                 EFDD   \n",
      "11         maritaulvskog  ...                  1                  S&D   \n",
      "12        MargotLJParker  ...                  6                 EFDD   \n",
      "13          Nigel_Farage  ...                  6                 EFDD   \n",
      "\n",
      "    retweetMepCountryId retweetMepCountryShort  \\\n",
      "3                    27                    GBR   \n",
      "5                    27                    GBR   \n",
      "11                   26                    SWE   \n",
      "12                   27                    GBR   \n",
      "13                   27                    GBR   \n",
      "\n",
      "                     origCreatedAt         origTweetId  \\\n",
      "3   Wed Jul 04 11:05:03 +0000 2012  220473289259233285   \n",
      "5   Fri Jul 05 08:00:23 +0000 2013  353060776707235841   \n",
      "11  Sat May 17 12:47:17 +0000 2014  467647542378123264   \n",
      "12  Wed May 28 16:47:44 +0000 2014  471694322472353793   \n",
      "13  Wed May 28 16:47:44 +0000 2014  471694322472353793   \n",
      "\n",
      "                  retweetCreatedAt      retweetTweetId lang  \\\n",
      "3   Sun Dec 07 21:46:00 +0000 2014  541710240014942209   en   \n",
      "5   Thu Jun 18 08:52:51 +0000 2015  611456527409524736   en   \n",
      "11  Wed Sep 16 09:15:06 +0000 2015  644077036022165504   en   \n",
      "12  Thu Dec 04 22:21:10 +0000 2014  540631925145493504   en   \n",
      "13  Sun Mar 29 18:53:32 +0000 2015  582254277122473984   en   \n",
      "\n",
      "                                                Tweet  \n",
      "3   Euro Parliament votes overwhelmingly to reject...  \n",
      "5   US have one currency, one Central Bank and one...  \n",
      "11  More jobs for young people is my top priority ...  \n",
      "12  Vote #Labour, get Tory. Vote Tory, get Labour....  \n",
      "13  Vote #Labour, get Tory. Vote Tory, get Labour....  \n",
      "\n",
      "[5 rows x 22 columns]\n"
     ]
    }
   ],
   "source": [
    "data_n = data\n",
    "\n",
    "data_n['Tweet'] = ''\n",
    "\n",
    "for tweet in filtered_tweets:\n",
    "    data_n.loc[data_n['origTweetId'] == tweet['id'], ['Tweet']] = tweet['text']\n",
    "\n",
    "print(data_n.head())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "After noticing that some of the Tweets contain a url which clearly means nothing for our model we can simply remove it using a regular expression."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "import re\n",
    "def removeUrl(inputTweet):\n",
    "    return re.sub(r\"http\\S+\", \"\", inputTweet, flags=re.MULTILINE)\n",
    "\n",
    "data_n['Tweet'] = data_n['Tweet'].apply(lambda x: removeUrl(x))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next, in order to clear our dataframe, we will remove the rows which have empty tweet text and we will print the length of it before and after the process in order to see whether our work have given any results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "13796\n",
      "12932\n"
     ]
    }
   ],
   "source": [
    "n_data_n = data_n\n",
    "print(len(n_data_n))\n",
    "n_data_n['Tweet'].replace('', np.nan, inplace=True)\n",
    "n_data_n.dropna(subset=['Tweet'], inplace=True)\n",
    "print(len(n_data_n))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Our next move is to group our results by the group of the MEP that posted each tweet."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<pandas.core.groupby.generic.DataFrameGroupBy object at 0x000001F798EA3898>\n"
     ]
    }
   ],
   "source": [
    "grouped = n_data_n.groupby('origMepGroupShort')\n",
    "\n",
    "print(grouped)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "scrolled": false
   },
   "source": [
    "Then we remove the groups which have less than 50 tweets and print the new length of the dataframe to see whether this process has made any difference in our dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "12908\n"
     ]
    }
   ],
   "source": [
    "final = grouped.filter(lambda x: len(x) > 50)\n",
    "print(len(final))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Clustering"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For this part of the exercise, we want cluster the tweets based on their text using k-means."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We start by using the count vectorizer provided from sklearn."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "CountVectorizer(analyzer='word', binary=False, decode_error='strict',\n",
       "                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\n",
       "                lowercase=True, max_df=1.0, max_features=None, min_df=1,\n",
       "                ngram_range=(1, 1), preprocessor=None, stop_words=None,\n",
       "                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\n",
       "                tokenizer=None, vocabulary=None)"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "\n",
    "vectorizer = CountVectorizer()\n",
    "vectorizer"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now using the CountVectorizer from above we will create a bag of words which we will use next for k-means. We convert everything to lowercase, use min_df=10, max_df=0.5 and remove all English stopwords."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0, 0, 0, ..., 0, 0, 0],\n",
       "       [0, 0, 0, ..., 0, 0, 0],\n",
       "       [0, 0, 0, ..., 0, 0, 0],\n",
       "       ...,\n",
       "       [0, 0, 0, ..., 0, 0, 0],\n",
       "       [0, 0, 0, ..., 0, 0, 0],\n",
       "       [0, 0, 0, ..., 0, 0, 0]], dtype=int64)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tf_vectorizer = CountVectorizer(strip_accents='unicode',\n",
    "                                lowercase=True,\n",
    "                                max_df=0.5, \n",
    "                                min_df=10,\n",
    "                                stop_words='english')\n",
    "\n",
    "X = tf_vectorizer.fit_transform(final['Tweet']).toarray()\n",
    "X"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next, we use the elbow method, to investigate the best number of clusters for our data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x2123680b898>"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "kmeans = KMeans(random_state=42)\n",
    "visualizer = KElbowVisualizer(kmeans, k=(1,11))\n",
    "\n",
    "visualizer.fit(X)\n",
    "visualizer.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The elbow method suggests using 6 clusters for our data but to be sure and validate this result we  can try the silhouette method in order to verify the results we got and settle on how many clusters we will use."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "For n_clusters = 2 The average silhouette_score is : 0.04911702862067595\n",
      "For n_clusters = 3 The average silhouette_score is : 0.04354153449388772\n",
      "For n_clusters = 4 The average silhouette_score is : 0.030265907840674428\n",
      "For n_clusters = 5 The average silhouette_score is : 0.030936327139231758\n",
      "For n_clusters = 6 The average silhouette_score is : 0.029887949595709465\n",
      "For n_clusters = 7 The average silhouette_score is : 0.021122910411837212\n",
      "For n_clusters = 8 The average silhouette_score is : 0.012934751461754435\n",
      "For n_clusters = 9 The average silhouette_score is : 0.013130592294198872\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x2880 with 8 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "plt.figure(figsize=(2 * 5,  10 * 4))\n",
    "silhouette_scores = []\n",
    "scores = {}\n",
    "for n_clusters in range(2, 10):\n",
    "    plt.subplot(10, 2, n_clusters - 1)\n",
    "    kmeans = KMeans(n_clusters, random_state=42)\n",
    "    visualizer = SilhouetteVisualizer(kmeans, colors='yellowbrick')\n",
    "    visualizer.fit(X)\n",
    "    scores[n_clusters] = visualizer.silhouette_score_\n",
    "    plt.title(f'clusters: {n_clusters} score: {visualizer.silhouette_score_}')\n",
    "    cluster_labels = kmeans.fit_predict(X)\n",
    "    silhouette_avg = silhouette_score(X, cluster_labels)\n",
    "    silhouette_scores.append(silhouette_avg)\n",
    "    print(\"For n_clusters =\", n_clusters,\n",
    "          \"The average silhouette_score is :\", silhouette_avg)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here we can see that silhouette method suggests using 2 clusters, but elbow method suggested 6 so we have to try another approach in order to decide the cluster number. In order to decide we will use the silhouette method scores in order to plot the along with their respected cluster sizes. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x243d0fa5320>]"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "size = range(2, 10)\n",
    "plt.plot(size, silhouette_scores)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here we can see that the score of the method keeps droping until the point of 4 and 6 clusters. These are the points where we could draw a vertical line indicating that this would be the point we would stop searching for the best cluster number. Since elbow method too suggested using 6 cluster we will try spliting our data into 6 clusters."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we will visualize the clusters using Yellowbrick's InterclusterDistance in order to have a clear view of how the data are clustered."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x243d1217668>"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model = KMeans(6, random_state=42)\n",
    "visualizer = InterclusterDistance(model)\n",
    "\n",
    "visualizer.fit(X)        \n",
    "visualizer.show() "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "From the last plot we can see that our model does a good job clustering these tweets into 6 clusters, creating only on mediocre conflict between clusters 4 and 5. Next we can print the centers of our clusters."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[-8.67361738e-18  4.69483568e-03  7.04225352e-03 ...  6.72205347e-18\n",
      "  -1.19262239e-18  7.04225352e-03]\n",
      " [-2.60208521e-18  1.70854271e-02  2.01005025e-03 ... -1.92987987e-17\n",
      "  -1.07336015e-17  1.00502513e-03]\n",
      " [ 2.09753540e-03  1.20608285e-02  6.29260619e-03 ...  1.57315155e-03\n",
      "   1.04876770e-03  2.09753540e-03]\n",
      " [ 8.47457627e-03  5.64971751e-03  1.12994350e-02 ...  6.07153217e-18\n",
      "   2.27682456e-18  5.85469173e-18]\n",
      " [ 2.89855072e-03  1.04084321e-02  5.66534914e-03 ...  1.05401845e-03\n",
      "   7.90513834e-04  1.71277997e-03]\n",
      " [ 1.83374083e-03  4.88997555e-03  4.88997555e-03 ...  1.22249389e-03\n",
      "   1.22249389e-03  1.22249389e-03]]\n"
     ]
    }
   ],
   "source": [
    "centers = model.cluster_centers_\n",
    "print(centers)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In order to find the top 30 features per cluster we have to find for each cluster centroid, which token vectors have the largest projection onto that centroid and then find the corresponding terms."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Cluster 0: meps labour eu amp today vote uk eurolabour support ttip action tory parliament meeting calling join voted energy steel help just work say saveoursteel ukip lnbdublin food cameron theprogressives debate\n",
      "------------------------------------------------------------------------------------\n",
      "Cluster 1: ukip vote labour today party nhs people amp eu uk poll just election policy lab immigration voters tories mep national support farage campaign britain great want tory new real voting\n",
      "------------------------------------------------------------------------------------\n",
      "Cluster 2: eu uk leave cameron referendum people britain today need european membership want just migration europe trade eppgroup turkey support vote policy borders says right member control campaign states immigration labour\n",
      "------------------------------------------------------------------------------------\n",
      "Cluster 3: debate eu amp ttip today vote european live parliament ep europarl_en digital votes plenary looking theprogressives need forward citizens aldegroup tonight greece ukip just want good key right come people\n",
      "------------------------------------------------------------------------------------\n",
      "Cluster 4: eppgroup today european europe great ep people meeting parliament good new europarl_en vote ttip support just greece need theprogressives uk work right thanks labour gt year day aldegroup digital refugees\n",
      "------------------------------------------------------------------------------------\n",
      "Cluster 5: amp eu eppgroup europe good european meeting today need people ttip ep uk migration great vote theprogressives support greece new rights parliament work jobs refugees trade eurolabour mep stop europarl_en\n",
      "------------------------------------------------------------------------------------\n"
     ]
    }
   ],
   "source": [
    "features = tf_vectorizer.get_feature_names()\n",
    "relevant_labels = list(set(model.labels_))\n",
    "centroids = model.cluster_centers_.argsort()[:,::-1]\n",
    "for this_label in relevant_labels:\n",
    "    print('Cluster {}:'.format(this_label), end='')\n",
    "    for ind in centroids[this_label, :30]:\n",
    "        print(' {}'.format(features[ind]), end='')\n",
    "    print('')\n",
    "    print('------------------------------------------------------------------------------------')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The words make sense because they are all possible and common words a member of a european political party would use. But with a quick look one could not really tell which cluster represents which party or another possible way in which the tweets could have been clustered."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Classification"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For the purpose of this exercise we want to train at least two algorithms to learn to classify an unseen tweet. The target variable should be the political party of the original poster and the training features should be the original tweet's text."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We start by spliting our data into a training set and a testing set. We choose to keep the 80% of our data for training our model and 20% of them for testing it. We also transform our categorical variable which described the political party in which the member of the parliament who made the Tweet belongs to a numerical one in order to use it next in our algorithms."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\pandas\\core\\generic.py:5208: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  self[name] = value\n"
     ]
    }
   ],
   "source": [
    "X = tf_vectorizer.fit_transform(final['Tweet']).toarray()\n",
    "tfidfconverter = TfidfTransformer()\n",
    "X = tfidfconverter.fit_transform(X).toarray()\n",
    "\n",
    "final.origMepGroupShort = pd.Categorical(final.origMepGroupShort)\n",
    "y = final[\"origMepGroupShort\"].cat.codes\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we will try various different algorithms for validating our data. We will use in each case cross validation in order to find the best parameters for the best algorithm for our case. <br>\n",
    "The first thing we will try is decision trees and more specifically random forest classification and extremely randomized trees in order to find how well the could fit our data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[206   8  35  43   1  16  54]\n",
      " [ 21 102  63  10   1   3  46]\n",
      " [ 16   5 545   5   0  10  39]\n",
      " [ 26   5  14 315   0  12  58]\n",
      " [  6   2   1   3  34   2  20]\n",
      " [ 27   4  20  15   0 113  46]\n",
      " [ 39   6  87  27   0  10 461]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.60      0.57      0.59       363\n",
      "           1       0.77      0.41      0.54       246\n",
      "           2       0.71      0.88      0.79       620\n",
      "           3       0.75      0.73      0.74       430\n",
      "           4       0.94      0.50      0.65        68\n",
      "           5       0.68      0.50      0.58       225\n",
      "           6       0.64      0.73      0.68       630\n",
      "\n",
      "    accuracy                           0.69      2582\n",
      "   macro avg       0.73      0.62      0.65      2582\n",
      "weighted avg       0.69      0.69      0.68      2582\n",
      "\n",
      "0.6878388845855926\n"
     ]
    }
   ],
   "source": [
    "classifier = RandomForestClassifier(n_estimators=100, random_state=0)\n",
    "classifier.fit(X_train, y_train) \n",
    "tree_pred = classifier.predict(X_test)\n",
    "\n",
    "print(confusion_matrix(y_test, tree_pred))\n",
    "print(classification_report(y_test, tree_pred))\n",
    "print(accuracy_score(y_test, tree_pred)) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "With accuracy of 0.687 we have found surely a good approch with our first try, but since this approach was the first, we really don't know how well we 've done, so it's important to keep trying other approaches."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[219   6  33  45   3  14  43]\n",
      " [ 18 109  64  13   0   4  38]\n",
      " [ 18   6 551   5   0  12  28]\n",
      " [ 23   6  13 322   1  14  51]\n",
      " [  3   1   5   5  36   2  16]\n",
      " [ 28   3  22  18   1 115  38]\n",
      " [ 35   7  75  34   0  17 462]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.64      0.60      0.62       363\n",
      "           1       0.79      0.44      0.57       246\n",
      "           2       0.72      0.89      0.80       620\n",
      "           3       0.73      0.75      0.74       430\n",
      "           4       0.88      0.53      0.66        68\n",
      "           5       0.65      0.51      0.57       225\n",
      "           6       0.68      0.73      0.71       630\n",
      "\n",
      "    accuracy                           0.70      2582\n",
      "   macro avg       0.73      0.64      0.67      2582\n",
      "weighted avg       0.71      0.70      0.70      2582\n",
      "\n",
      "0.7025561580170411\n"
     ]
    }
   ],
   "source": [
    "extremely_rts = ExtraTreesClassifier(n_estimators=50, \n",
    "                                     max_depth=None,\n",
    "                                     min_samples_split=2, \n",
    "                                     random_state=42)\n",
    "\n",
    "scores = cross_val_score(extremely_rts, X_train, y_train, cv=5)\n",
    "extremely_rts.fit(X_train, y_train)\n",
    "extra_tree_pred = extremely_rts.predict(X_test)\n",
    "\n",
    "print(confusion_matrix(y_test, extra_tree_pred))\n",
    "print(classification_report(y_test, extra_tree_pred))\n",
    "print(accuracy_score(y_test, extra_tree_pred)) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This time with extra trees classifier we have done even better with average score of 0.702! We keep trying other approaches in order to find the one which fits in our data best."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[189   0  53  52   1  12  56]\n",
      " [ 18  50 109  21   0   0  48]\n",
      " [  6   0 563  11   0   2  38]\n",
      " [ 29   2  24 302   0  10  63]\n",
      " [  7   0  14   5  18   2  22]\n",
      " [ 28   1  44  30   0  75  47]\n",
      " [ 37   2  79  30   0   8 474]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.60      0.52      0.56       363\n",
      "           1       0.91      0.20      0.33       246\n",
      "           2       0.64      0.91      0.75       620\n",
      "           3       0.67      0.70      0.69       430\n",
      "           4       0.95      0.26      0.41        68\n",
      "           5       0.69      0.33      0.45       225\n",
      "           6       0.63      0.75      0.69       630\n",
      "\n",
      "    accuracy                           0.65      2582\n",
      "   macro avg       0.73      0.53      0.55      2582\n",
      "weighted avg       0.67      0.65      0.62      2582\n",
      "\n",
      "0.6471727343144849\n"
     ]
    }
   ],
   "source": [
    "text_clf = Pipeline([\n",
    "                     ('clf', MultinomialNB())\n",
    "                    ])\n",
    "\n",
    "text_clf.fit(X_train, y_train)\n",
    "multinom_pred = text_clf.predict(X_test)\n",
    "print(confusion_matrix(y_test, multinom_pred))\n",
    "print(classification_report(y_test, multinom_pred))\n",
    "print(accuracy_score(y_test, multinom_pred)) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The multinomial distribution gave a score of 0.647 but we have to set a barrier in order to know how good that was. Stil not better than ExtraTreesClassifier."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[156  47  17  39  17  74  13]\n",
      " [ 19 150  28   9  10  19  11]\n",
      " [ 25 159 322  10  12  78  14]\n",
      " [ 54  47   8 218   9  72  22]\n",
      " [  9   5   3   6  31   9   5]\n",
      " [ 25  17   9  17  12 137   8]\n",
      " [ 61  82  56  58  16 107 250]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.45      0.43      0.44       363\n",
      "           1       0.30      0.61      0.40       246\n",
      "           2       0.73      0.52      0.61       620\n",
      "           3       0.61      0.51      0.55       430\n",
      "           4       0.29      0.46      0.35        68\n",
      "           5       0.28      0.61      0.38       225\n",
      "           6       0.77      0.40      0.52       630\n",
      "\n",
      "    accuracy                           0.49      2582\n",
      "   macro avg       0.49      0.50      0.47      2582\n",
      "weighted avg       0.59      0.49      0.51      2582\n",
      "\n",
      "0.4895429899302866\n"
     ]
    }
   ],
   "source": [
    "gauss = GaussianNB()\n",
    "scores = cross_val_score(gauss, X_train, y_train, cv=5)\n",
    "gauss.fit(X_train, y_train)\n",
    "gaus_pred = gauss.predict(X_test)\n",
    "print(confusion_matrix(y_test, gaus_pred))\n",
    "print(classification_report(y_test, gaus_pred))\n",
    "print(accuracy_score(y_test, gaus_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The last approach using Naive Bayes method proved below average so based on every other result we got so far it is really not so good for our case. <br>\n",
    "Finally we will try the DummyClassifier in order to set a baseline for our scores."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[  0   0 363   0   0   0   0]\n",
      " [  0   0 246   0   0   0   0]\n",
      " [  0   0 620   0   0   0   0]\n",
      " [  0   0 430   0   0   0   0]\n",
      " [  0   0  68   0   0   0   0]\n",
      " [  0   0 225   0   0   0   0]\n",
      " [  0   0 630   0   0   0   0]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.00      0.00      0.00       363\n",
      "           1       0.00      0.00      0.00       246\n",
      "           2       0.24      1.00      0.39       620\n",
      "           3       0.00      0.00      0.00       430\n",
      "           4       0.00      0.00      0.00        68\n",
      "           5       0.00      0.00      0.00       225\n",
      "           6       0.00      0.00      0.00       630\n",
      "\n",
      "    accuracy                           0.24      2582\n",
      "   macro avg       0.03      0.14      0.06      2582\n",
      "weighted avg       0.06      0.24      0.09      2582\n",
      "\n",
      "0.24012393493415957\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1268: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    }
   ],
   "source": [
    "dummy = DummyClassifier(strategy='most_frequent',random_state=42)\n",
    "scores = cross_val_score(dummy, X_train, y_train, cv=5)\n",
    "dummy.fit(X_train, y_train)\n",
    "dummy_pred = dummy.predict(X_test)\n",
    "print(classification_report(y_test, dummy_pred))\n",
    "print(accuracy_score(y_test, dummy_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "From the last score we can see that all our approaches did good and everyone of them at least doubled the score of the DummyClassifier who gave an accuracy score of 0.24. One of them stood out of the crowd and that was ExtraTreesClassifier which is the best approach for our case, succeeding in classification with 70% accuracy. <br>\n",
    "Now we can try to find the best parameters for the best classifier we found in order to see if we can do even better."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 81 candidates, totalling 243 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=3)]: Using backend LokyBackend with 3 concurrent workers.\n",
      "[Parallel(n_jobs=3)]: Done  35 tasks      | elapsed:    4.4s\n",
      "[Parallel(n_jobs=3)]: Done 156 tasks      | elapsed:  3.7min\n",
      "[Parallel(n_jobs=3)]: Done 243 out of 243 | elapsed:  6.8min finished\n"
     ]
    }
   ],
   "source": [
    "model = ExtraTreesClassifier(class_weight='balanced',\n",
    "                             n_estimators=50, \n",
    "                             max_depth=None,\n",
    "                             min_samples_split=2, \n",
    "                             random_state=42)\n",
    "\n",
    "parameters = {'criterion': ['entropy'],\n",
    "                       'min_samples_split' : [1, 2, 3],\n",
    "                       'max_depth' : [15,20,25],\n",
    "                      'n_estimators': [200,300,400],\n",
    "                      'max_features': ['log2'],\n",
    "                      'min_samples_leaf':range(0,3)\n",
    "             }\n",
    "\n",
    "\n",
    "clf = GridSearchCV(model, parameters, verbose=2, scoring='accuracy',\n",
    "                        cv=3, n_jobs=3)\n",
    "\n",
    "clf.fit(X_train,y_train)\n",
    "\n",
    "optimised_random_forest = clf.best_estimator_\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[213  21  25  33  18  41  12]\n",
      " [ 17 126  51  11   4  20  17]\n",
      " [ 22  30 520   3   4  29  12]\n",
      " [ 31  18  11 307   9  38  16]\n",
      " [  5   1   4   1  51   4   2]\n",
      " [ 19  12  19  19   3 138  15]\n",
      " [ 52  33  69  21  11  41 403]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.59      0.59      0.59       363\n",
      "           1       0.52      0.51      0.52       246\n",
      "           2       0.74      0.84      0.79       620\n",
      "           3       0.78      0.71      0.74       430\n",
      "           4       0.51      0.75      0.61        68\n",
      "           5       0.44      0.61      0.51       225\n",
      "           6       0.84      0.64      0.73       630\n",
      "\n",
      "    accuracy                           0.68      2582\n",
      "   macro avg       0.63      0.66      0.64      2582\n",
      "weighted avg       0.70      0.68      0.68      2582\n",
      "\n",
      "0.6808675445391169\n"
     ]
    }
   ],
   "source": [
    "final_pred = optimised_random_forest.predict(X_test)\n",
    "print(confusion_matrix(y_test, final_pred))\n",
    "print(classification_report(y_test, final_pred))\n",
    "print(accuracy_score(y_test, final_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "ExtraTreesClassifier(bootstrap=False, ccp_alpha=0.0, class_weight='balanced',\n",
       "                     criterion='entropy', max_depth=20, max_features='log2',\n",
       "                     max_leaf_nodes=None, max_samples=None,\n",
       "                     min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "                     min_samples_leaf=1, min_samples_split=2,\n",
       "                     min_weight_fraction_leaf=0.0, n_estimators=200,\n",
       "                     n_jobs=None, oob_score=False, random_state=42, verbose=0,\n",
       "                     warm_start=False)"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "optimised_random_forest"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Even though we seem to have found the best estimator given the parameters we provided, the score of 68% is lower and not higher than 70% which we expected to find."
   ]
  }
 ],
 "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}