Air Line Twitter Sentiment Analysis.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/matplotlib-2.2.2+772.g640ecd9-py3.5-linux-x86_64.egg/matplotlib/__init__.py:961: UserWarning: could not find rc file; returning defaults\n",
      "  warnings.warn(message)\n"
     ]
    }
   ],
   "source": [
    "import numpy as np \n",
    "import pandas as pd \n",
    "pd.set_option('display.max_colwidth', -1)\n",
    "from time import time\n",
    "import re\n",
    "import string\n",
    "import os\n",
    "import emoji\n",
    "from pprint import pprint\n",
    "import collections\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "sns.set(style=\"darkgrid\")\n",
    "sns.set(font_scale=1.3)\n",
    "\n",
    "from sklearn.base import BaseEstimator, TransformerMixin\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.pipeline import Pipeline, FeatureUnion\n",
    "from sklearn.metrics import classification_report\n",
    "\n",
    "from sklearn.naive_bayes import MultinomialNB\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.externals import joblib\n",
    "'''\n",
    "import gensim\n",
    "\n",
    "from nltk.corpus import stopwords\n",
    "from nltk.stem import PorterStemmer\n",
    "from nltk.tokenize import word_tokenize\n",
    "'''\n",
    "from nltk.corpus import stopwords\n",
    "from nltk.stem import PorterStemmer\n",
    "from nltk.tokenize import word_tokenize\n",
    "import gensim"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "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>text</th>\n",
       "      <th>airline_sentiment</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>@VirginAmerica  for all my flight stuff wrong and did nothing about it. Had #worst #flight ever</td>\n",
       "      <td>negative</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>@southwestair yeah, 7am flight tomorrow, going to add 60-90 min extra because of the roads / snow</td>\n",
       "      <td>neutral</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>@united no worries Your customer service gets a bad wrap but just spoke w agent who saved me huge amounts of time &amp;amp; apologized for yesterday</td>\n",
       "      <td>positive</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>@United @Skywest do it again!   I will now miss one or probably both our Aspen events. #Fail</td>\n",
       "      <td>negative</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>@USAirways 4469, thank you for asking.</td>\n",
       "      <td>neutral</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                                                                                                               text  \\\n",
       "0  @VirginAmerica  for all my flight stuff wrong and did nothing about it. Had #worst #flight ever                                                    \n",
       "1  @southwestair yeah, 7am flight tomorrow, going to add 60-90 min extra because of the roads / snow                                                  \n",
       "2  @united no worries Your customer service gets a bad wrap but just spoke w agent who saved me huge amounts of time &amp; apologized for yesterday   \n",
       "3  @United @Skywest do it again!   I will now miss one or probably both our Aspen events. #Fail                                                       \n",
       "4  @USAirways 4469, thank you for asking.                                                                                                             \n",
       "\n",
       "  airline_sentiment  \n",
       "0  negative          \n",
       "1  neutral           \n",
       "2  positive          \n",
       "3  negative          \n",
       "4  neutral           "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_csv('Tweets.csv')\n",
    "df = df.reindex(np.random.permutation(df.index)) \n",
    "df.reset_index(inplace=True)\n",
    "df.drop('index',inplace=True,axis=1)\n",
    "df = df[['text', 'airline_sentiment']]\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "class TextCounts(BaseEstimator, TransformerMixin):\n",
    "    \n",
    "    def count_regex(self, pattern, tweet):\n",
    "        #finding all the substring containing the pattern in the tweet\n",
    "        return len(re.findall(pattern, tweet))\n",
    "    \n",
    "    def fit(self, X, y=None, **fit_params):\n",
    "        # fit method is used when specific operations need to be done on the train data, but not on the test data\n",
    "        return self\n",
    "    \n",
    "    def transform(self, X, **transform_params):\n",
    "        #all the alphanumeric character\n",
    "        count_words = X.apply(lambda x: self.count_regex(r'\\w+', x)) \n",
    "        count_mentions = X.apply(lambda x: self.count_regex(r'@\\w+', x))\n",
    "        count_hashtags = X.apply(lambda x: self.count_regex(r'#\\w+', x))\n",
    "        count_capital_words = X.apply(lambda x: self.count_regex(r'\\b[A-Z]{2,}\\b', x))\n",
    "        count_excl_quest_marks = X.apply(lambda x: self.count_regex(r'!|\\?+', x))\n",
    "        count_urls = X.apply(lambda x: self.count_regex(r'https?://[^\\s]+[\\s]?', x))\n",
    "        # We will replace the emoji symbols with a description, which makes using a regex for counting easier\n",
    "        # Moreover, it will result in having more words in the tweet\n",
    "        count_emojis = X.apply(lambda x: emoji.demojize(x)).apply(lambda x: self.count_regex(r':[a-z_&]+:', x))\n",
    "        \n",
    "        df = pd.DataFrame({'count_words': count_words\n",
    "                           , 'count_mentions': count_mentions\n",
    "                           , 'count_hashtags': count_hashtags\n",
    "                           , 'count_capital_words': count_capital_words\n",
    "                           , 'count_excl_quest_marks': count_excl_quest_marks\n",
    "                           , 'count_urls': count_urls\n",
    "                           , 'count_emojis': count_emojis\n",
    "                          })\n",
    "        \n",
    "        return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "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>count_capital_words</th>\n",
       "      <th>count_emojis</th>\n",
       "      <th>count_excl_quest_marks</th>\n",
       "      <th>count_hashtags</th>\n",
       "      <th>count_mentions</th>\n",
       "      <th>count_urls</th>\n",
       "      <th>count_words</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>26</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   count_capital_words  count_emojis  count_excl_quest_marks  count_hashtags  \\\n",
       "0  0                    0             0                       2                \n",
       "1  0                    0             0                       0                \n",
       "2  0                    0             0                       0                \n",
       "3  0                    0             1                       1                \n",
       "4  0                    0             0                       0                \n",
       "\n",
       "   count_mentions  count_urls  count_words  \n",
       "0  1               0           16           \n",
       "1  1               0           17           \n",
       "2  1               0           26           \n",
       "3  2               0           17           \n",
       "4  1               0           6            "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "\n",
    "tc = TextCounts()\n",
    "\n",
    "df_eda = tc.fit_transform(df.text)\n",
    "df_eda.head()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "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>count_capital_words</th>\n",
       "      <th>count_emojis</th>\n",
       "      <th>count_excl_quest_marks</th>\n",
       "      <th>count_hashtags</th>\n",
       "      <th>count_mentions</th>\n",
       "      <th>count_urls</th>\n",
       "      <th>count_words</th>\n",
       "      <th>airline_sentiment</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>16</td>\n",
       "      <td>negative</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>17</td>\n",
       "      <td>neutral</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>26</td>\n",
       "      <td>positive</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>17</td>\n",
       "      <td>negative</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "      <td>neutral</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   count_capital_words  count_emojis  count_excl_quest_marks  count_hashtags  \\\n",
       "0  0                    0             0                       2                \n",
       "1  0                    0             0                       0                \n",
       "2  0                    0             0                       0                \n",
       "3  0                    0             1                       1                \n",
       "4  0                    0             0                       0                \n",
       "\n",
       "   count_mentions  count_urls  count_words airline_sentiment  \n",
       "0  1               0           16           negative          \n",
       "1  1               0           17           neutral           \n",
       "2  1               0           26           positive          \n",
       "3  2               0           17           negative          \n",
       "4  1               0           6            neutral           "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_eda['airline_sentiment'] = df.airline_sentiment\n",
    "df_eda.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def show_dist(df, col):\n",
    "    print('Descriptive stats for {}'.format(col))\n",
    "    print('-'*(len(col)+22))\n",
    "    print(df.groupby('airline_sentiment')[col].describe())\n",
    "    bins = np.arange(df[col].min(), df[col].max() + 1)\n",
    "    g = sns.FacetGrid(df, col='airline_sentiment', size=5, hue='airline_sentiment', palette=\"PuBuGn_d\")\n",
    "    g = g.map(sns.distplot, col, kde=False, norm_hist=True, bins=bins)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Descriptive stats for count_words\n",
      "---------------------------------\n",
      "                    count       mean       std  min   25%   50%   75%   max\n",
      "airline_sentiment                                                          \n",
      "negative           9178.0  20.387557  6.100020  2.0  17.0  22.0  25.0  35.0\n",
      "neutral            3099.0  15.290094  7.150738  2.0  10.0  15.0  21.0  34.0\n",
      "positive           2363.0  14.478206  7.266129  2.0  8.0   14.0  21.0  33.0\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/matplotlib-2.2.2+772.g640ecd9-py3.5-linux-x86_64.egg/matplotlib/axes/_axes.py:6178: UserWarning: The 'normed' kwarg is deprecated, and has been replaced by the 'density' kwarg.\n",
      "  warnings.warn(\"The 'normed' kwarg is deprecated, and has been \"\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABCQAAAFUCAYAAAD8jO1vAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAgIhjIAEAAAAAACLuH8T2+nCJ3uotAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 1080x360 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "show_dist(df_eda, 'count_words')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Descriptive stats for count_mentions\n",
      "------------------------------------\n",
      "                    count      mean       std  min  25%  50%  75%  max\n",
      "airline_sentiment                                                     \n",
      "negative           9178.0  1.111244  0.365420  1.0  1.0  1.0  1.0  6.0\n",
      "neutral            3099.0  1.167473  0.480102  1.0  1.0  1.0  1.0  5.0\n",
      "positive           2363.0  1.138383  0.432462  1.0  1.0  1.0  1.0  6.0\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/matplotlib-2.2.2+772.g640ecd9-py3.5-linux-x86_64.egg/matplotlib/axes/_axes.py:6178: UserWarning: The 'normed' kwarg is deprecated, and has been replaced by the 'density' kwarg.\n",
      "  warnings.warn(\"The 'normed' kwarg is deprecated, and has been \"\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABCQAAAFUCAYAAAD8jO1vAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nYK+rHKcAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 1080x360 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_dist(df_eda, 'count_mentions')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Descriptive stats for count_hashtags\n",
      "------------------------------------\n",
      "                    count      mean       std  min  25%  50%  75%  max\n",
      "airline_sentiment                                                     \n",
      "negative           9178.0  0.233384  0.615099  0.0  0.0  0.0  0.0  6.0\n",
      "neutral            3099.0  0.212004  0.666382  0.0  0.0  0.0  0.0  6.0\n",
      "positive           2363.0  0.297503  0.774776  0.0  0.0  0.0  0.0  8.0\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/matplotlib-2.2.2+772.g640ecd9-py3.5-linux-x86_64.egg/matplotlib/axes/_axes.py:6178: UserWarning: The 'normed' kwarg is deprecated, and has been replaced by the 'density' kwarg.\n",
      "  warnings.warn(\"The 'normed' kwarg is deprecated, and has been \"\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABCQAAAFUCAYAAAD8jO1vAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nANiOQAIAAAAAANiOQAIAAAAAANju/wEHA98vctNtmwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 1080x360 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_dist(df_eda, 'count_hashtags')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Descriptive stats for count_capital_words\n",
      "-----------------------------------------\n",
      "                    count      mean       std  min  25%  50%  75%   max\n",
      "airline_sentiment                                                      \n",
      "negative           9178.0  0.429070  1.019710  0.0  0.0  0.0  1.0  22.0\n",
      "neutral            3099.0  0.462085  0.939468  0.0  0.0  0.0  1.0  11.0\n",
      "positive           2363.0  0.461278  1.234078  0.0  0.0  0.0  1.0  20.0\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/matplotlib-2.2.2+772.g640ecd9-py3.5-linux-x86_64.egg/matplotlib/axes/_axes.py:6178: UserWarning: The 'normed' kwarg is deprecated, and has been replaced by the 'density' kwarg.\n",
      "  warnings.warn(\"The 'normed' kwarg is deprecated, and has been \"\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABCQAAAFUCAYAAAD8jO1vAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAAADYjoYEAAAAAACw3Q90Z68gsu24WQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 1080x360 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_dist(df_eda, 'count_capital_words')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Descriptive stats for count_excl_quest_marks\n",
      "--------------------------------------------\n",
      "                    count      mean       std  min  25%  50%  75%   max\n",
      "airline_sentiment                                                      \n",
      "negative           9178.0  0.589235  1.041339  0.0  0.0  0.0  1.0  27.0\n",
      "neutral            3099.0  0.639238  0.868104  0.0  0.0  0.0  1.0  12.0\n",
      "positive           2363.0  0.890394  1.198117  0.0  0.0  1.0  1.0  13.0\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/matplotlib-2.2.2+772.g640ecd9-py3.5-linux-x86_64.egg/matplotlib/axes/_axes.py:6178: UserWarning: The 'normed' kwarg is deprecated, and has been replaced by the 'density' kwarg.\n",
      "  warnings.warn(\"The 'normed' kwarg is deprecated, and has been \"\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABCQAAAFUCAYAAAD8jO1vAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAAEZ0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yKzc3\nMi5nNjQwZWNkOSwgaHR0cDovL21hdHBsb3RsaWIub3JnL0J0boQAACAASURBVHic7N17WJR1/v/x\nF4FmOpCSZOupLGU08YBkYLKe8pQH0C41LU+pa9lXKk0F7WDl7rpGqatttaJZZtZKmaGdy0o0dbPM\nUkNXUiFYldSCMQXFz+8Pf846gjLAcMOMz8d1eV3OPZ+Zeb/nnnmDL++5x88YYwQAAAAAAGChKyq7\nAAAAAAAAcPkhkAAAAAAAAJYjkAAAAAAAAJYjkAAAAAAAAJYjkAAAAAAAAJYjkAAAAAAAAJYjkPAS\n3bp106JFi0pcZ7fb9d577130si/bsmWL7Ha7cnJyKrsUFONyei3CNzB3S8bc9U6X02sUVQ+ztWTM\n1orlzmtp1apVatWqlUUVXd4IJLzEW2+9peHDh5f6dhs2bFD37t0roKLK1apVK61atcplW3h4uDZs\n2KBrrrmmkqpyNXbsWCUkJFR2GZa7WN+++lqE72LuumLuVo733ntPdru9sssAPIbZ6orZar3zX0s5\nOTmy2+3asmWLy5o+ffro888/r4zyLjsBlV0A3BMcHHzJ6wsKClS9evUi20NCQiqqpCqnevXql1W/\n3oZ9A2/D3C0Zc7fquNjrEahqmK0lY7ZWLHee2xo1aqhGjRoWVAOOkKgCNm7cqBEjRujWW29VRESE\nhg8fru+//95lzYWHt9ntdi1fvlwPPfSQwsPD9cQTTxR73+cfkvTzzz/Lbrfrww8/1Pjx49WmTRt1\n795da9ascbnN4cOHNWXKFEVGRioiIkIjRozQzp073erl1KlT+vOf/6w//vGPCgsLU6dOnfTnP//Z\neb0xRq+88op69uypVq1aqXfv3lq2bJmMMS41v/HGG3rkkUcUHh6uzp076+WXX3Z5LgoKCjR9+nTZ\n7Xbn/xxdeHjbuUOtNm3apH79+ql169YaMWKEDh06pK+//loDBgxQ27ZtNWrUKB06dMilj/Xr12vw\n4MFq3bq1unTpoqeffloOh8N5/YgRI/TEE0/o+eef12233abIyEjNmDFDJ06ckCQlJCRow4YNeued\nd5w1Xpi8esq5vjdt2qRhw4apdevW6tevn7766iuXdfv27dOECRMUERGhyMhI3X///crIyHBZs3jx\nYkVHR6tt27Z68MEH9frrr7v8z1xmZqYmTpyo6OhotWnTRv3799fq1aud11+q7/Nfi0OHDtWTTz7p\n8thnzpxRp06dtHjxYknuvVaAsmLuMnfLw1Nzd+HCherdu3ex952Tk6MtW7Zo8uTJkuTs6dz/Uo4Y\nMUKPP/64nnvuOXXs2NF5P2vWrNHgwYOdjzl+/Hjt27evQp4H4ELMVmZreZzr+4svvtCdd96pVq1a\nqX///kUe75tvvtHQoUPVqlUrRUZG6tFHH3Xpaffu3Ro9erQiIiIUHh6u/v3768svv3Ref/5rKTo6\nWpI0cuRI2e12devWTZLrRzb27dsnu91e5LW8YcMG3Xzzzc79VJ7X22XNoNJ9/PHH5r333jPp6elm\nz5495rHHHjPt27c3R48eda7p2rWr+ec//+m8HBoaam699VazfPlyc+DAAXPgwAHn9rVr17qsO3c5\nMzPThIaGmttvv9188MEHZv/+/eaZZ54xLVu2NJmZmcYYY06cOGF69eplJk2aZHbs2GF++uknM2fO\nHBMREWEOHz5cYi9LliwxnTp1Mlu2bDFZWVlm27Zt5s0333ReP3/+fNOtWzfz2WefmYyMDPPxxx+b\nyMhI8/rrr7vU3LFjR7Ny5Uqzf/9+88orr5jQ0FCzdetWY4wxR44cMWFhYeaVV14xhw8fdta1efNm\nExoa6rz89ttvm+bNm5vhw4eb7777zuzYscP07NnTDBs2zAwfPtx8++235scffzR33HGHiYuLcz7+\nxo0bTZs2bcyKFSvM/v37zTfffGPuvPNO89BDDznXDB8+3ERERJi//e1vJj093Xz55ZcmIiLCvPDC\nC8YYY3Jzc82IESPMQw895KwxPz+/2Ofs3XffNW3btr3knxdffPGiz/m5vmNjY01qaqrZt2+fmTp1\nqmnfvr3Jy8szxhhz+PBhExUVZWbNmmV2795t/vOf/5ipU6earl27mhMnThhjjHn//fdNy5YtzfLl\ny81PP/1kli9fbm699VYTGhrqfKy0tDTz2muvmR9//NEcOHDALF++3LRo0cJs2rSpxL7Pfy2uWLHC\n3HrrraagoMB531999ZVp3ry5OXjwoNuvFaCsmLvM3aowdxcsWGB69epV7H2fq/+NN95wXj58+LDJ\nzc11Ph9t27Y1Tz31lNm7d6/ZvXu3McaYt956y3z22WfmwIEDZufOnWbChAmmR48eLs/Fha9ZwFOY\nrcxWT8zWnj17mi+++ML85z//MQkJCaZt27bml19+McYYc+jQIdO2bVsTHx9v9uzZYzZu3Gi6detm\nHnzwQef99OvXzzzyyCNm79695sCBA+bTTz81W7Zscdkv515LaWlpJjQ01Hz00Ufm8OHD5siRI87n\nPCwszHmbwYMHm1mzZrnUO2XKFDN27FhjTPlfb5czAokq6MyZMyYyMtK8++67zm3FDe/HH3+8yG3d\nGd6vvvqq8/pTp06Z1q1bm1WrVhljzv4i07VrV3P69GmX++3Tp49JSkoqsfZZs2aZkSNHmjNnzhS5\n7vfffzetW7c2X331lcv2pKQk06dPH5ea//rXv7qs6dGjh1m4cKHzclhYmHn77bdd1hQ3vENDQ52/\npBljzNKlS01oaKj54YcfnNuWLVtmbrnlFufle+65x8ybN8/lvrdv325CQ0OdQ2r48OFmwIABLmse\nffRRM3z4cOflMWPGmPj4+CLPw4Xy8vLM/v37L/nn2LFjF739ub4//fRT57aDBw+a0NBQ5/D9+9//\nboYNG+Zyu1OnTpmIiAjz/vvvG2OMueuuu0xCQoLLmvj4eJdAojgTJ040jz76aIl9n/9a/PXXX03L\nli3NJ5984vJY9957rzHG/dcK4CnMXeZuZczdkgIJY4xZu3ZtsXN4+PDhpnfv3sXu9wt7tdvtzn8A\nGUMgAeswW5mtZZmt5/ahMcYUFBSYP/7xj+Yf//iHMcaYuXPnmq5du5pTp04516xfv96EhoY6w6zw\n8HCX+7jQ+a+lw4cPm9DQULN582aXNRcGEsuXLzdRUVHOxz1+/Lhp27atSUlJMcaU//V2OeMcElVA\nZmamFixYoO+++05HjhyRMUYnTpxQdnb2JW9X1jO/tmjRwvn3gIAABQcH65dffpEk7dixQwcPHtQt\nt9zicpv8/Pwih/cXZ+DAgRozZox69uypjh07qlOnTurcubP8/f31n//8RydPntSECRPk5+fnvM3p\n06eLHIZ/fo2SdO211zprLI2AgAA1bdrUeblu3bqS5PIxhLp16yo3N9f5mcUdO3Zo+/btevXVV51r\nztV34MAB52cfmzdvXqTGb7/9ttQ12mw22Wy2Ut/uQuc/Z9dee60kuezX7du3Kzw83OU2J06ccO7X\n9PR0DRw40OX6tm3b6p133nFZ/49//EOff/65cnJydOrUKRUUFCgyMrJUtV599dXq2rWrUlJS1L17\nd504cUIfffSR82McpXmtAGXB3GXuVoW5W15hYWEu+1WSfvzxRz3//PP68ccfdezYMUlnn8vs7GxF\nRER45HGBi2G2Mls9MVvbtm3r/Hu1atXUqlUrpaenSzr7+2rbtm0VEPC/f8a2b99ekrR37141btxY\n9957rx577DG98847ioyMVK9evVyeu7Lo06ePZs+erdTUVHXt2lUff/yx/Pz81KNHD0nlf71dzggk\nqoD7779fwcHBmjlzpq677jpVq1ZNw4cP16lTpy55u5o1a5bp8apVq+Zy2c/Pzzmczpw5o2bNmmnB\nggVFbufOgGnZsqU+++wzbdy4UZs2bdKMGTN04403unym7vnnn1ejRo1KXeOZM2dKfPwLXXHFFbri\niv+dKuXcD43z7//ctvOfg/vvv1/9+vUrcn/16tXzeI0pKSmaOXPmJdfcd999uv/++y+5pqSeoqOj\nNWPGjCK3u/rqq92u9ZlnntHnn3+uadOm6cYbb9RVV12lxMRE/fbbb27fxzmxsbGaNGmScnNznZ/r\nOzfUS/NaAcqCuetejczdip27578Ozjl9+nTJDfx/V111lcvlEydOaMyYMWrfvr1mz56tunXrys/P\nT3379i3xtQ14ArPVvRqZrZeereUVFxenmJgYrV+/Xhs3btQLL7ygGTNm6J577inzfdapU0edO3fW\nu+++6/xPtV69ejlPfFne19vljECikh07dkx79+7Vyy+/rI4dO0qSDh06VKbk1BPCwsK0Zs0aBQUF\nqU6dOmW6D5vNpl69eqlXr1666667NGDAAO3Zs0dNmzbVlVdeqaysLOcJZMqqWrVqKiwsLNd9XEzL\nli31008/6frrry/X/bhbY7du3dSmTZtLrilNaFCcli1bas2aNfrDH/5w0bOw33TTTdq+fbvuuusu\n57bt27e7rNm6datiY2PVp08fSWeH7759+1zOmO1u3507d1bNmjX1wQcf6JNPPlHPnj2dv5B48rUC\nXIi5WzbM3dJxZ+5ec801OnLkiM6cOeP8h8aFJ0A79w+FwsJC+fv7X/Ix09PTdfToUU2ePFk33HCD\npLNzvCz/sABKi9laNszWorZv364mTZpIOnty0R9++MH5++lNN92kNWvW6PTp086jJL7++mtJcjkK\n4vrrr9eIESM0YsQI/eUvf9G//vWvYgOJczPWnTkZGxurKVOmKD09XZs2bdLSpUud13ni9Xa5IpCo\nZFdffbWCg4P1r3/9S/Xr19evv/6qxMTESvuamf79+2vp0qWaMGGCHnroITVq1Eg5OTlKTU3Vbbfd\nVuQwpAstWbJE9erVU4sWLVStWjWlpKToqquuUv369VWrVi2NHz9eiYmJMsYoKipK+fn52rVrlw4f\nPqz77rvP7TobNmyoLVu2qFOnTqpWrVqJXyFVGg8++KDGjRun6667Tv3791eNGjX0008/6fPPP9es\nWbNKVeNXX32ljIwM2Ww2BQYGFkmgJc8d3nYpI0aM0Ntvv624uDjdd999CgkJ0X//+1999tlnGjZs\nmG644QaNGjVKU6dOVatWrdShQwd99dVXWrduncv9NGnSRJ9++qluv/121axZU0uXLtXBgwddnn93\n+65WrZr69u2r5cuXKz093fntGpI8+loBLsTcZe5WlbkbGRmp48ePa8GCBbrzzju1Y8cOvfnmm0V6\nkqR169YpIiJCV155pWrVqlXsY9avX1/Vq1fXa6+9ptGjR+vnn39WYmKiy/+qAhWF2cps9dRsfeml\nl1SnTh01bNhQS5YsUW5urjOQuOeee7Rs2TI9/vjjGjt2rHJycvTkk0+qd+/eaty4sY4fP665c+eq\nZ8+eatCggY4ePaqtW7fqpptuKvaxateurcDAQG3YsEFNmzZV9erVLxqadOnSRVdddZUmT56sevXq\nuXxkubyvt8sZP6Eq2RVXXKG///3v2r9/v2JiYpSQkKDhw4e7HEZlpRo1amj58uVq1qyZpk6dqt69\ne2vSpEnav3+/8/Oxl1KzZk0tWbJEgwYNUmxsrLZt26aXXnpJtWvXliRNnDhRU6ZM0euvv65+/fpp\n5MiRSk5Odv7C5a74+Hjt2rVLt99+uzp06FCmXi+mQ4cOevnll/X9999r6NChGjhwoObNm1fq74Me\nPXq0goODFRsbqw4dOpTps3ieUrduXb3xxhuqWbOmJkyYoD59+mjGjBlyOBwKCgqSJN1xxx16+OGH\n9Y9//EOxsbHauHGj/vSnP+nKK6903s/06dN13XXXaeTIkRo1apTq1q2rvn37ujxWafqOjY3Vnj17\nVLduXUVFRblc56nXCnAh5i5z1wruzN0bb7xRs2bNUkpKivNrlCdNmuRyPzfffLPuvfdezZw5U7fd\ndtsl/xERHBysxMREpaamqk+fPvrb3/6madOmuXzWGqgozFZmq6dMmzZN8+fPV2xsrL7//nu9+OKL\nznNmXHvttVq8eLH27dungQMH6uGHH1ZUVJT+8pe/SDp7ro1ff/1V06dPV+/evTVhwgQ1b978kh8l\neeKJJ/Thhx+qS5cuRc6ndr7q1aurd+/eSktLU0xMjMv5Q8r7eruc+RnOEAfgIp566ilt3bq1yPd6\nAwAAAJ60ZcsWjRw5Uhs2bCh1aALvRWQOQNLZz+gtXbpUnTt31pVXXqnU1FQlJydr2rRplV0aAAAA\nAB9EIIFSGTdunL755ptir+vfv7+efvppiyuCp/j5+WnLli1asmSJTpw4ocaNGyshIaFcZyQGUH7M\nXQDwPGYrUDXwkQ2UyqFDh3Ty5Mlir7PZbLrmmmssrggAfBtzFwA8j9kKVA0EEgAAAAAAwHJV6iMb\nOTl5Zb5tnTo1dezY7x6spuqgN+/ly/3Rm7VCQgLLfR/M2OLRm/fy5f7ozVrM2Irjy71Jvt0fvXmn\nqtrbxeasz3ztZ0CAf2WXUGHozXv5cn/0dnnx5eeE3ryXL/dHb5cXX35OfLk3ybf7ozfv5G29+Uwg\nAQAAAAAAvAeBBAAAAAAAsByBBAAAAAAAsByBBAAAAAAAsByBBAAAAAAAsByBBAAAAAAAsByBBAAA\nAAAAsByBBAAAAAAAsByBBAAAAAAAsByBBAAAAAAAsByBBAAAAAAAsFxAZRfgKclfbJIj76Rba/tF\n3VLB1QAAAAAAgEvhCAkAAAAAAGA5AgkAAAAAAGA5twKJwsJCzZkzR1FRUQoPD1dcXJyOHj160fVH\njhxRfHy8IiMj1a5dO8XGxurQoUMeKxoAAAAAAHg3twKJRYsWad26dUpOTtb69eslSdOmTSt2bX5+\nvkaPHq1q1arpgw8+0NatW/Xss8+qVq1anqsaAAAAAAB4NbcCiZUrV2rcuHFq1KiRAgMDNXXqVKWm\npiorK6vI2nfeeUe5ubmaOXOmgoODdcUVV6hZs2ay2WweLx4AAAAAAHgnP2OMudSC3NxctW/fXqtX\nr1aLFi2c2yMiIvTMM8/o9ttvd1k/adIkHTlyRCEhIdqwYYPq1KmjoUOHavTo0SUWc/p0oQIC/MvU\nSPIXm9xeO7hLhzI9BgB4s/LMWADApTFjAaD0Svzaz+PHj0tSkSMcgoKC5HA4iqw/duyYtmzZohkz\nZmj27NnavXu3xo0bp+DgYMXExFzysY4d+700tRfh7td+5uTkletxrBYSEuh1NbvLl3uTfLs/erNW\nSEhgue+jPDO2Kj4nnkJv3suX+6M3azFjK44v9yb5dn/05p2qam8Xm7MlfmTj3LkfLgwfcnNzi/0Y\nRq1atVSvXj2NGjVK1atXV6tWrRQTE6PPPvusLHUDAAAAAAAfVGIgERQUpPr162vnzp3ObRkZGXI4\nHLLb7UXWt2jRQn5+fkW2F7cNAAAAAABcntw6qeWQIUOUlJSkzMxM5eXlKTExUdHR0WrYsGGRtQMH\nDtSvv/6q119/XYWFhUpLS9OaNWvUs2dPjxcPAAAAAAC8k1uBxPjx49WtWzcNGjRInTp1kjFGiYmJ\nkqSUlBSFh4c71zZo0ECLFi1ScnKyIiIi9OCDDyouLk59+vSpmA4AAAAAAIDXKfGklpLk7++v+Ph4\nxcfHF7kuJiamyMkqIyMjtXr1as9UCAAAAAAAfI5bR0gAAAAAAAB4EoEEAAAAAACwHIEEAAAAAACw\nHIEEAAAAAACwHIEEAAAAAACwHIEEAAAAAACwHIEEAAAAAACwHIEEAAAAAACwHIEEAAAAAACwHIEE\nAAAAAACwHIEEAAAAAACwHIEEAAAAAACwHIEEAAAAAACwHIEEAAAAAACwHIEEAAAAAACwHIEEAAAA\nAACwHIEEAAAAAACwHIEEAAAAAACwHIEEAAAAAACwHIEEAAAAAACwHIEEAAAAAACwHIEEAAAAAACw\nHIEEAAAAAACwHIEEAAAAAACwHIEEAAAAAACwHIEEAAAAAACwHIEEAAAAAACwHIEEAAAAAACwXEBl\nFwAAqPre3L5PDsdJt9bGhjas4GoAAADgCzhCAgAAAAAAWI5AAgAAAAAAWI5AAgAAAAAAWM6tQKKw\nsFBz5sxRVFSUwsPDFRcXp6NHjxa7dsuWLbLb7QoPD3f+GTp0qEeLBgAAAAAA3s2tk1ouWrRI69at\nU3JysmrXrq0ZM2Zo2rRpWrx4cbHr/f39tW3bNo8WCgAAAAAAfIdbR0isXLlS48aNU6NGjRQYGKip\nU6cqNTVVWVlZFV0fAAAAAADwQSUeIZGbm6vs7GyFhYU5tzVu3Fg2m01paWlq0KBBkdsUFhaqc+fO\nOn36tFq2bKnJkyerefPmJRZTp05NBQT4l7KF/7EF1nBrXUhIYJkfo7J4Y83u8uXeJN/uj968S7lm\nbPYvstmYsd7Il3uTfLs/evMu5f091hefk3N8uTfJt/ujN+/kTb2VGEgcP35ckmSz2Vy2BwUFyeFw\nFFl/44036t1331XTpk31+++/KykpSaNGjVJKSorq1at3ycc6duz30tRehCPvpFvrcnLyyvU4VgsJ\nCfS6mt3ly71Jvt0fvVnLEz9Yyj1jHcxYb+PLvUm+3R+9WauyZ2xVfE48xZd7k3y7P3rzTlW1t4vN\n2RI/slGrVi1JKhI+5ObmFgkpzj5QiJo3b66AgAAFBQXpkUceUe3atbV+/fqy1A0AAAAAAHxQiYFE\nUFCQ6tevr507dzq3ZWRkyOFwyG63V2hxAAAAAADAN7l1UsshQ4YoKSlJmZmZysvLU2JioqKjo9Ww\nYcMiazdt2qQDBw7ozJkzOn78uBYuXKgjR44oOjra48UDAAAAAADv5FYgMX78eHXr1k2DBg1Sp06d\nZIxRYmKiJCklJUXh4eHOtbt379bo0aPVrl07de/eXd99951efvll/eEPf6iYDgAAAAAAgNcp8aSW\nkuTv76/4+HjFx8cXuS4mJkYxMTHOy6NHj9bo0aM9ViAAAAAAAPA9bh0hAQAAAAAA4EkEEgAAAAAA\nwHIEEgAAAAAAwHIEEgAAAAAAwHIEEgAAAAAAwHIEEgAAAAAAwHIEEgAAAAAAwHIEEgAAAAAAwHIE\nEgAAAAAAwHIBlV0AAAAAcDnbmH5IjryTbq1t1yC4gqsBAOtwhAQAAAAAALAcgQQAAAAAALAcgQQA\nAAAAALAcgQQAAAAAALAcgQQAAAAAALAcgQQAAAAAALAcgQQAAAAAALAcgQQAAAAAALAcgQQAAAAA\nALAcgQQAAAAAALAcgQQAAAAAALAcgQQAAAAAALAcgQQAAAAAALAcgQQAAAAAALAcgQQAAAAAALAc\ngQQAAAAAALAcgQQAAAAAALAcgQQAAAAAALAcgQQAAAAAALAcgQQAAAAAALAcgQQAAAAAALAcgQQA\nAAAAALCcW4FEYWGh5syZo6ioKIWHhysuLk5Hjx4t8XYrVqyQ3W7XCy+8UO5CAQAAAACA73ArkFi0\naJHWrVun5ORkrV+/XpI0bdq0S94mKytLS5cuVWhoaPmrBAAAAAAAPsWtQGLlypUaN26cGjVqpMDA\nQE2dOlWpqanKysq66G0effRRTZo0SbVr1/ZYsQAAAAAAwDcElLQgNzdX2dnZCgsLc25r3LixbDab\n0tLS1KBBgyK3efPNN3XVVVepT58+euONN9wupk6dmgoI8Hd7/YVsgTXcWhcSEljmx6gs3lizu3y5\nN8m3+6M371KuGZv9i2w2Zqw38uXeJN/uj968S3lm7J7c3/k91ov5cn/05p28qbcSA4njx49Lkmw2\nm8v2oKAgORyOIuuzs7P14osvauXKlaUu5tix30t9m/M58k66tS4nJ69cj2O1kJBAr6vZXb7cm+Tb\n/dGbtTzxg6XcM9bBjPU2vtyb5Nv90Zu1qsSM5fdYr+TL/dGbd6qqvV1szpb4kY1atWpJUpHwITc3\nt0hIIUmPPfaYJkyYoHr16pWlTgAAAAAAcBko8QiJoKAg1a9fXzt37lSLFi0kSRkZGXI4HLLb7UXW\nb9y4UTt37tS8efMknQ0yfvjhB23YsEErVqzwcPkAAAAAAMAblRhISNKQIUOUlJSkyMhI1a5dW4mJ\niYqOjlbDhg2LrP3yyy9dLj/00EOKiIjQmDFjPFMxAAAAAADwem59y8b48ePVrVs3DRo0SJ06dZIx\nRomJiZKklJQUhYeHO9ded911Ln+qV68um82munXrVkwHAAAAAADA67h1hIS/v7/i4+MVHx9f5LqY\nmBjFxMRc9LavvfZa2asDAAAAAAA+ya0jJAAAAAAAADyJQAIAAAAAAFiOQAIAAAAAAFjOrXNIAADg\nrnf3/Oz22tjQot/WBAAAgMsDR0gAAAAAAADLEUgAAAAAAADLEUgAAAAAAADLEUgAAAAAAADLEUgA\nAAAAAADLEUgAAAAAAADLEUgAAAAAAADLEUgAAAAAAADLEUgAAAAAAADLEUgAAAAAAADLEUgAAAAA\nAADLEUgAAAAAAADLEUgAAAAAAADLEUgAAAAAAADLEUgAAAAAAADLEUgAAAAAAADLEUgAAAAAAADL\nBVR2AQAAAADc823WUbfXtmsQXIGVAED5cYQEAAAAAACwHIEEAAAAAACwHIEEAAAAAACwHIEEAAAA\nAACwHIEEAAAAAACwHIEEAAAAAACwHIEEAAAAAACwHIEEAAAAAACwHIEEAAAAAACwHIEEAAAAAACw\nnFuBRGFhoebMmaOoqCiFh4crLi5OR48eLXbt1q1bNXDgQN16662KiIjQwIED9fHHH3u0aAAAAAAA\n4N3cCiQWLVqkdevWKTk5WevXr5ckTZs2rdi1TZo00fPPP68tW7Zo69atmjFjhqZOnar09HTPVQ0A\nAAAAALyaW4HEypUrNW7cODVq1EiBgYGaOnWqUlNTlZWVVWTtNddcowYNGsjPz0/GGPn5+enMmTM6\ncOCAx4sHAAAAAADeKaCkBbm5ucrOzlZYWJhzW+PGjWWz2ZSWlqYGDRoUe7tbbrlFJ06c0OnTp9W+\nfXtFR0eXWEydOjUVEOBfivJd2QJruLUuJCSwzI9R5hziBAAAHw9JREFUWbyxZnf5cm+Sb/dHb96l\nXDM2+xfZbO7N2NKoKs9zVamjIvhyb5Jv90dv3qU8M3ZP7u9u/x5bGlXlea4qdVQUX+6P3ryTN/VW\nYiBx/PhxSZLNZnPZHhQUJIfDcdHbbd26VQUFBVq/fr1++ukn+fuXPKCPHfu9xDWX4sg76da6nJy8\ncj2O1UJCAr2uZnf5cm+Sb/dHb9byxA+Wcs9Yh3sztjSqwvNcFfe3p/hyb5Jv90dv1qoSM9bN32NL\noyo8z1Vxf3uSL/dHb96pqvZ2sTlb4kc2atWqJUlFwofc3NwiIcWFqlevru7du+vrr79WcnKyu7UC\nAAAAAAAfV2IgERQUpPr162vnzp3ObRkZGXI4HLLb7W49SGFhIeeQAAAAAAAATm6d1HLIkCFKSkpS\nZmam8vLylJiYqOjoaDVs2LDI2o8++ki7d+/W6dOnlZ+fr5UrV2rz5s1unUMCAAAAAABcHko8h4Qk\njR8/Xrm5uRo0aJAKCgrUsWNHJSYmSpJSUlI0c+ZMbdu2TZKUk5Oj5557Tjk5OapWrZqaNGmi5557\nTh07dqy4LgAAAAAAgFdxK5Dw9/dXfHy84uPji1wXExOjmJgY5+Xhw4dr+PDhnqsQAAAAAAD4HLc+\nsgEAAAAAAOBJBBIAAAAAAMByBBIAAAAAAMByBBIAAAAAAMByBBIAAAAAAMBybn3Lhq9Zu3mr22v7\nRd1SgZUAAAAAAHB54ggJAAAAAABgOQIJAAAAAABgOQIJAAAAAABgOQIJAAAAAABgOQIJAAAAAABg\nOQIJAAAAAABgOQIJAAAAAABgOQIJAAAAAABgOQIJAAAAAABgOQIJAAAAAABgOQIJAAAAAABgOQIJ\nAAAAAABgOQIJAAAAAABgOQIJAAAAAABgOQIJAAAAAABgOQIJAAAAAABgOQIJAAAAAABgOQIJAAAA\nAABgOQIJAAAAAABgOQIJAAAAAABgOQIJAAAAAABgOQIJAAAAAABgOQIJAAAAAABgOQIJAAAAAABg\nOQIJAAAAAABgOQIJAAAAAABgObcCicLCQs2ZM0dRUVEKDw9XXFycjh49WuzaL7/8UiNHjlRkZKTa\nt2+vu+++W1u3bvVo0QAAAAAAwLu5FUgsWrRI69atU3JystavXy9JmjZtWrFrf/vtN40YMUKffPKJ\nNm3apH79+ulPf/qT/vvf/3quagAAAAAA4NXcCiRWrlypcePGqVGjRgoMDNTUqVOVmpqqrKysImtj\nYmLUo0cPBQUFKSAgQHfffbdq1qypH374wePFAwAAAAAA71RiIJGbm6vs7GyFhYU5tzVu3Fg2m01p\naWklPsDu3bt17NgxhYaGlq9SAAAAAADgMwJKWnD8+HFJks1mc9keFBQkh8NxydseOXJEDz74oMaM\nGaMbbrihxGLq1KmpgAD/EtddjC2wRplvezEhIYEev8+yqCp1VARf7k3y7f7ozbuUa8Zm/yKbjRnr\njXy5N8m3+6M371KeGbsn93d+j/VivtwfvXknb+qtxECiVq1aklQkfMjNzS0SUpzv0KFDGjNmjDp2\n7KhHHnnErWKOHfvdrXUX48g7Wa7bFycnJ8/j91laISGBVaKOiuDLvUm+3R+9WcsTP1jKPWMdzFhv\n48u9Sb7dH71Zq0rMWH6P9Uq+3B+9eaeq2tvF5myJH9kICgpS/fr1tXPnTue2jIwMORwO2e32Ym/z\n888/65577lGnTp30xBNPyM/Pr4xlAwAAAAAAX+TWSS2HDBmipKQkZWZmKi8vT4mJiYqOjlbDhg2L\nrE1PT9fdd9+tvn37Kj4+3uMFAwAAAAAA7+dWIDF+/Hh169ZNgwYNUqdOnWSMUWJioiQpJSVF4eHh\nzrWLFy/WoUOHtGzZMoWHhzv/pKSkVEwHAAAAAADA65R4DglJ8vf3V3x8fLFHPMTExCgmJsZ5efbs\n2Zo9e7bnKgQAAAAAAD7HrSMkAAAAAAAAPIlAAgAAAAAAWI5AAgAAAAAAWI5AAgAAAAAAWI5AAgAA\nAAAAWI5AAgAAAAAAWI5AAgAAAAAAWI5AAgAAAAAAWI5AAgAAAAAAWI5AAgAAAAAAWI5AAgAAAAAA\nWI5AAgAAAAAAWI5AAgAAAAAAWI5AAgAAAAAAWI5AAgAAAAAAWI5AAgAAAAAAWI5AAgAAAAAAWI5A\nAgAAAAAAWI5AAgAAAAAAWI5AAgAAAAAAWI5AAgAAAAAAWI5AAgAAAAAAWI5AAgAAAAAAWI5AAgAA\nAAAAWI5AAgAAAAAAWI5AAgAAAAAAWC6gsgsAAAAA4HnfZh11e227BsEVWAkAFI8jJAAAAAAAgOUI\nJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOXcCiQKCws1Z84cRUVFKTw8XHFxcTp6\ntPiz9h46dEgTJkxQ165dZbfb9e6773q0YAAAAAAA4P3cCiQWLVqkdevWKTk5WevXr5ckTZs2rdi1\nfn5+io6O1rPPPqvrrrvOc5UCAAAAAACfEeDOopUrV+qBBx5Qo0aNJElTp05Vjx49lJWVpQYNGris\nvfbaa3XPPfdIkvz9/T1cLgAAAAAA8AUlHiGRm5ur7OxshYWFObc1btxYNptNaWlpFVocAAAAAADw\nTSUeIXH8+HFJks1mc9keFBQkh8Ph0WLq1KmpgICyH1VhC6zhwWrOCgkJ9Ph9lkVVqaMi+HJvkm/3\nR2/epVwzNvsX2WzMWG/ky71Jvt0fvXmX8szYPbm/V8jvsaVRkfvEF/f3+Xy5P3rzTt7UW4mBRK1a\ntSSpSPiQm5tbJKQor2PHfi/X7R15Jz1Uyf/k5OR5/D5LKyQksErUURF8uTfJt/ujN2t54gdLuWes\ngxnrbXy5N8m3+6M3a1WJGVsBv8eWRkXtk6q4vz3Jl/ujN+9UVXu72Jwt8SMbQUFBql+/vnbu3Onc\nlpGRIYfDIbvd7rkKAQAAAADAZcOtb9kYMmSIkpKSlJmZqby8PCUmJio6OloNGzYsdn1+fr7y8/Nl\njNHp06eVn5+v06dPe7RwAAAAAADgvdwKJMaPH69u3bpp0KBB6tSpk4wxSkxMlCSlpKQoPDzcZX3r\n1q3VunVrZWdna8aMGWrdurVefPFFz1cPAAAAAAC8kltf++nv76/4+HjFx8cXuS4mJkYxMTEu23bv\n3u2Z6gAAAAAAgE9y6wgJAAAAAAAATyKQAAAAAAAAliOQAAAAAAAAliOQAAAAAAAAlnPrpJYAAFSE\nd/f87Pba2NDiv2oaAAAA3okjJAAAAAAAgOU4QqIEazdvdXttv6hbKrASAAAAAAB8B0dIAAAAAAAA\nyxFIAAAAAAAAyxFIAAAAAAAAyxFIAAAAAAAAyxFIAAAAAAAAyxFIAAAAAAAAyxFIAAAAAAAAyxFI\nAAAAAAAAyxFIAAAAAAAAyxFIAAAAAAAAywVUdgEAAAAAKte3WUfdXtuuQXAFVgLgcsIREgAAAAAA\nwHIEEgAAAAAAwHIEEgAAAAAAwHIEEgAAAAAAwHIEEgAAAAAAwHIEEgAAAAAAwHIEEgAAAAAAwHIB\nlV2AL1m7eavba/tF3VKBlQAAAAAAULVxhAQAAAAAALAcR0gAALzCu3t+dnttbGjDCqwEAC5v32Yd\ndXttr5DACqwEgLfjCAkAAAAAAGA5AgkAAAAAAGA5AgkAAAAAAGA5AgkAAAAAAGA5t05qWVhYqGef\nfVbvvPOO8vPzFR0draeeekrBwcHFrl+/fr3mzJmjzMxMNW7cWAkJCYqOjvZo4d6uNF8Rem//rhVY\nCQAAAAAA1nMrkFi0aJHWrVun5ORk1a5dWzNmzNC0adO0ePHiImszMzMVFxenp59+WnfccYc+/PBD\nTZw4UWvXrlXDhpz1HAAAAEBRpfn2jnYNiv+PUQDexa1AYuXKlXrggQfUqFEjSdLUqVPVo0cPZWVl\nqUGDBi5r33nnHbVs2VKxsbGSpJiYGL355ptavXq1Jk6c6OHyLw/JX2ySI++kW2v7Rd1SwdUAQNVX\nmq8ItWXXkMPh3ozl60QBoHQ2ph9y+/dYAJefEgOJ3NxcZWdnKywszLmtcePGstlsSktLKxJIpKWl\nqWXLli7bbr75ZqWlpXmoZFxKaT4KUhXwcRQA3qQ0QUdpEHQAQOmU5miKimTL/d3twKWijurgyBJ4\nsxIDiePHj0uSbDaby/agoCA5HI5i1wcGBhZZu3fv3hKLCQkJLHHNxQzu0qHMt0XlKs9+9wa+3B+9\neZfy9DTUB5+Py4UvvpbP58v90Zt3KU9Pvvh8wDq9KvD148uvTXqrGkr8lo1atWpJUpHwITc3t0hI\ncW59Xl6eW2sBAAAAAMDlqcRAIigoSPXr19fOnTud2zIyMuRwOGS324usb968uXbt2uWybdeuXWre\nvLkHygUAAAAAAL6gxEBCkoYMGaKkpCRlZmYqLy9PiYmJio6OLvZbMwYMGKAdO3Zo7dq1KigoUEpK\ninbt2qUBAwZ4vHgAAAAAAOCd/IwxpqRFhYWFevbZZ7Vq1SoVFBSoY8eOevrppxUcHKyUlBTNnDlT\n27Ztc65fv3695syZo8zMTDVq1EjTp09XdHR0hTYCAAAAAAC8h1uBBAAAAAAAgCe59ZENAAAAAAAA\nTyKQAAAAAAAAlvPqQKKwsFBz5sxRVFSUwsPDFRcXp6NHj1Z2WR6RkJCgli1bKjw83Pnn9ddfr+yy\nyuS9997T3XffrXbt2unmm28ucv3q1avVvXt3tWnTRoMHD9aOHTsqocqyu1R/q1atUvPmzV324+TJ\nkyup0tJLTExU37591a5dO0VHR+uxxx7Tr7/+6rLGW/dfSb15+77zFF+ds740YyXfnrPMWO/cd8xY\n9zBjvQMz1jvfp8xYL9l3xou98MILpmfPniYjI8Pk5uaaiRMnmrFjx1Z2WR4RHx9vZsyYUdlleMT6\n9evNmjVrTHJysmnRooXLdV9//bVp06aNSU1NNfn5+WbRokWmQ4cOJi8vr5KqLb1L9ff222+b7t27\nV1Jl5ffcc8+ZnTt3moKCAnPkyBEzduxYc9999zmv9+b9V1Jv3r7vPMVX56wvzVhjfHvOMmO9c98x\nY93DjPUOzFjvxIz1jn3n1UdIrFy5UuPGjVOjRo0UGBioqVOnKjU1VVlZWZVdGs7zxz/+Uf369VOj\nRo2KXJecnKwePXooOjpa1atX17hx43TllVfqk08+qYRKy+ZS/Xm7yZMn6+abb1a1atUUHByskSNH\n6t///rfzem/efyX1hrOYs97Bl+csM9Y79x0z1j3MWO/AjPVOzFjv4LWBRG5urrKzsxUWFubc1rhx\nY9lsNqWlpVViZZ7z8ccf69Zbb1WvXr00Z84cHT9+vLJL8ri0tDS1bNnSednPz0/Nmzf3mX0oSf/9\n73/VsWNHde7cWZMmTVJmZmZll1RmmzZtUvPmzZ2XfWn/Xdib5Fv7rix8fc5eDjNW8q33aXF86X3K\njPXefVcWzFjf4Evv0+L40vuUGVs1953XBhLnhprNZnPZHhQUJIfDURkledTw4cP1wQcfaPPmzXr+\n+ef19ddf6/HHH6/ssjzu+PHjCgwMdNnmK/tQktq3b681a9YoNTVVb731lq688kqNGTNGv//+e2WX\nVmofffSR3nzzTT366KPObb6y/4rrzZf2XVn58py9XGas5Dvv0+L40vuUGeu9+66smLG+wVfep8Xx\npfcpM7bq7juvDSRq1aolSUVeMLm5uUUGuzcKCwtT3bp1dcUVV6hZs2aaMWOGPvroIxUUFFR2aR5V\nq1Yt5eXluWzzlX0oSY0aNVKTJk10xRVXKCQkRLNmzVJOTo62b99e2aWVygcffKDHH39cL774okuS\n7Av772K9+cq+Kw9fnrOXy4yVfON9ejG+8j5lxnrvvisPZqxv8IX36cX4yvuUGVu1953XBhJBQUGq\nX7++du7c6dyWkZEhh8Mhu91eiZVVDD8/P0mSMaaSK/Gs5s2ba9euXc7Lxhj9+OOPRQ458hXeuB/f\nfvttzZw5Uy+++KKioqJcrvP2/Xep3i7kjfuuvC6nOevL+9fb36el4Y37kRl7ljfuu/JixvoGb3+f\nloY37kdm7FlVed95bSAhSUOGDFFSUpIyMzOVl5enxMRERUdHq2HDhpVdWrm99957ys3NlSTt379f\nc+bMUbdu3XTllVdWcmWlV1hYqPz8fJ06dUqSlJ+fr/z8fBljNHjwYH3yySfatGmTCgoKtHjxYhUU\nFKhHjx6VXLX7LtXfF198oYMHD8oYo19//VVPPfWU6tSpozZt2lRy1e5ZtmyZnnnmGS1evFgRERFF\nrvfm/VdSb96+7zzFV+esL81YybfnLDPWO/cdM9Y9zFjvwIz1zvcpM9Y79p2fqYoxiZsKCwv17LPP\natWqVSooKFDHjh319NNPKzg4uLJLK7cRI0Zo9+7dKigoUHBwsHr06KG4uDivOoTonFWrVmn69OlF\ntn/22Wdq2LChVq9erYULFyonJ0ehoaF68sknXU7wVNVdqr/XX39da9askcPhkM1mU7t27TRp0iQ1\nadKkEiotPbvdroCAAFWvXt1l+7Zt25x/99b9V1Jvc+bM8ep95ym+Omd9acZKvj1nmbHeue+Yse5h\nxnoHZqx3vk+Zsd6x77w6kAAAAAAAAN7Jqz+yAQAAAAAAvBOBBAAAAAAAsByBBAAAAAAAsByBBAAA\nAAAAsByBBAAAAAAAsByBBAAAAAAAsByBBAAAAAAAsByBBHzali1bZLfblZOTU9ml+JyEhASNHTu2\nsssAUImYsRWHGQtAYs5WJOZs1UAgAY/IycmR3W7Xli1bKrsUn9C7d28tXLiwsssAUEUwYz2LGQvg\nQsxZz2LOwl0BlV0AAO9SWFgoPz+/yi4DAHwSMxYAKhZztmrhCAkfYIzRsmXLdMcddygsLEwdO3ZU\nQkKCJCkvL08zZsxQZGSkWrVqpaFDh2rbtm3O217sMDC73a733ntPkvTzzz/Lbrfrww8/1Pjx49Wm\nTRt1795da9asca6Pjo6WJI0cOVJ2u13dunVzq/bvv/9eo0aNUtu2bdWxY0dNmTJFR44ckSQVFBQo\nNjZWDz/8sHO9w+HQ7bffrieffNK5bf/+/YqLi1P79u3Vpk0bDRgwQJs3by7FM/g/GzduVN++fdWq\nVSvdeeed+uKLL1zScneeL0k6fPiwpkyZosjISEVERGjEiBHauXOn8/q8vDxNmzZNHTp0UKtWrdSt\nWzctWrRIkjRixAjt27dPzz//vOx2u+x2u37++edL1n1uH7333nu699571aZNG/Xu3Vv//ve/dejQ\nIf3pT39S27Zt1adPH23dutV5O2OMHnvsMXXv3l2tW7fW7bffrrlz56qgoMC5ZuHCherdu7dSUlLU\nq1cvtWrVStnZ2UVqOHDggLp3767p06fr9OnTl+wR8CbMWGYsMxaoWMxZ5ixz9vJFIOED5s+fr/nz\n52vUqFFau3atkpKS1KxZM0nS9OnT9fXXX2v+/PlatWqVGjdurLFjx+qXX34p9eM8++yzuvPOO51v\n5unTpzuHS0pKiqSzb/gNGzborbfeKvH+9uzZo1GjRikqKkqrVq1SUlKSjh07pgceeEDGGFWvXl1z\n587VF198oeTkZEnSzJkzVbNmTU2fPl3S2WE5bNgwnThxQosWLdLatWs1ceLEUvcmSYcOHdKECRPU\nrl07rV69WnFxcfrrX/9a6vs5efKkRo4cqTNnzujll1/WW2+9pVatWmnUqFHO4T9//nzt3r1b//zn\nP/Xhhx/qb3/7m/7whz9IOvscXn/99RozZow2bNigDRs2OK8ryfz583XPPfdo9erVatasmSZPnqz4\n+HgNGTJEq1evVmhoqCZPnuwc0sYYXXPNNXruuef0/vvv69FHH9U777yjl156yeV+Dx48qOTkZCUm\nJmrNmjW65pprXK7/4YcfNGzYMN1xxx2aPXu2AgICLtkj4E2YsczYc5ixQMVgzjJnz2HOXoYMvJrD\n4TCtWrUyr776apHr9u3bZ0JDQ83GjRud2woKCkyXLl3MvHnzjDHGbN682YSGhprDhw+73DY0NNSs\nXbvWGGNMZmamCQ0NdXmMU6dOmdatW5tVq1YZY4w5fPiwCQ0NNZs3b3a79mnTppkpU6a4bDty5IgJ\nDQ0127dvd25buXKladOmjZk7d65p3bq12bNnj/O6uXPnmujoaHPixIliH+Ni/RVn7ty5plu3bub0\n6dPObcnJyS59ufN8vfXWW6Zr164u92OMMX369DFJSUnGGGPuu+8+k5CQcNFaevXqZRYsWFBizeec\n20evvfaac9vOnTtNaGioWbJkiXNbWlqaCQ0NNT/++ONF72vZsmWmR48ezssLFiwwzZs3NwcPHnRZ\nFx8fb8aMGWNSU1NNu3btirwGS+oR8AbMWGasMcxYoCIxZ5mzxjBnL2ecQ8LLpaenKz8/Xx07diz2\nOj8/P0VERDi3VatWTW3btlV6enqpH6tFixbOvwcEBCg4OLhM6fQ5O3bs0IEDB/Tpp58WuS4jI0Ot\nW7eWJA0ePFiff/65XnrpJT3xxBPOxFySdu3apYiICNWoUaPMdZyzd+9etWnTRv7+/s5t7dq1K/X9\n7NixQwcPHtQtt9zisj0/P18ZGRmSpGHDhumhhx7Sjh07dNttt6lLly7q0KFD+RqQ1Lx5c+ff69at\nK+nsIXjnhISESJLLflu5cqWSk5OVlZWlEydO6PTp0zLGuNzvtddeq3r16hV5vLS0NN1///2aOXOm\nBg8e7HJdRfUIWIkZy4w9HzMW8DzmLHP2fMzZyw+BxGXuiivOfmrn/DftqVOnil1brVo1l8t+fn5F\n3uylcebMGd15553Fft3O+YdR5eXlKS0tTf7+/tq/f3+ZH88T3Hm+zpw5o2bNmmnBggVFbm+z2SRJ\nnTt31rp165SamqrNmzfrgQceUJcuXTRv3rxy1RcQ8L+39LmT9Zy/7fwaJemDDz7Q008/rSlTpuiW\nW26RzWbTxx9/rOeee85l/VVXXVXs4zVo0ED16tXTu+++qz59+qhWrVrO6yqqR8CbMGNLhxnrihkL\nlIw5WzrMWVfM2crHOSS83E033aTq1atr48aNRa5r2rSpjDH65ptvnNtOnTql7777Tk2bNpUkBQcH\nSzr7+bVzdu3aVeo6zg34c8PBHS1bttSePXt0/fXXF/lzbthJ0uOPPy6bzaakpCStWLFCX375pfO6\nm2++Wd9++61OnjxZ6pov1LRpU23fvl2FhYXObeefNEly7/kKCwtTZmamgoKCivR1/g+n4OBgxcbG\navbs2UpMTNT777+v3377TdLZ5/P8OirK1q1bFRYWptGjRyssLEw33HCDsrKy3L59YGCgXnnlFZ08\neVLjxo2Tw+Fwuf5SPQLegBnLjC0PZixQMuYsc7Y8mLPej0DCy9WqVUv33nuv5s+frzfffFMHDhzQ\nrl27tGTJEl1//fXq2bOnZs6cqU2bNmnv3r169NFH9dtvv+mee+6RJF1//fVq0KCBFi5cqPT0dG3d\nulVz5swpdR21a9dWYGCgNmzYoJycHLfeqPfdd592796t6dOna8eOHcrMzNTGjRv12GOPOYfBW2+9\npc8//1xz585Vx44ddd999ykhIcF5Qp27775b+fn5mjhxorZt26bMzEx9+umnZToz8bBhw3T48GE9\n9dRTSk9P15dfflnkTLruPF/9+/fXddddpwkTJmjTpk36+eeftW3bNi1YsMB5VuB58+bp008/1f79\n+5Wenq6PPvpIISEhCgoKkiQ1bNhQ3377rbKzs3X06NFS/XAsjSZNmigtLU3r1q1TRkaGXn31VX34\n4Yeluo+goCAtXbpUZ86ccRnkJfUIeANmLDO2PJixQMmYs8zZ8mDOej8CCR8wadIk/d///Z9efvll\n9e3bV+PHj9fevXslSX/961/Vvn17Pfzwwxo4cKAyMjK0ZMkS52eyAgICNG/ePOXk5GjgwIGaNWuW\nHnnkkTLV8cQTT+jDDz9Uly5dNHDgwBLXN2vWTCtWrNDRo0c1atQo9evXT7NmzVKNGjVUvXp17du3\nT3/5y1+UkJDgTMH/7//+T9dff70SEhJkjFG9evW0YsUKBQQEaOzYserfv79eeOGFMn238HXXXacX\nXnhBW7duVWxsrObPn+/8yqlz3Hm+atSooeXLl6tZs2aaOnWqevfurUmTJmn//v269tprJZ1NjefN\nm6cBAwZo6NChOnLkiBYvXuysOy4uTg6HQ71791aHDh2K/WoiT7jrrrsUExOjhIQEDRgwQN99950e\nfPDBUt9PYGCglixZIj8/P917773Kzc0tsUfAWzBjmbFlxYwF3MOcZc6WFXPW+/mZ8nxwCvBxOTk5\nio6O1rJlyxQZGVnZ5QCAT2HGAkDF+n/t3bGNhDAQBVCHBEjQC2UgukCCKgiogy6oiYCADgguOmlX\na4KT2LnV6r3UyUQ/GNnfcpZP54YEAAAAEM4vG7xN27aX17OGYUjjOIbO0/f9UynSo67r0jzPofP8\nxTRNaV3X7FnTNGlZluCJgP8mY+8jY4EcOXsfOcsVTzZ4m23b0nme2bOqqlJd16Hz7Pt+2WBcluVT\na/CnOY7jpfX3V1EU2X+Vge8mY+8jY4EcOXsfOcsVCwkAAAAgnA4JAAAAIJyFBAAAABDOQgIAAAAI\nZyEBAAAAhPsBkj/ZzinovRcAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 1080x360 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_dist(df_eda, 'count_excl_quest_marks')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Descriptive stats for count_urls\n",
      "--------------------------------\n",
      "                    count      mean       std  min  25%  50%  75%  max\n",
      "airline_sentiment                                                     \n",
      "negative           9178.0  0.049248  0.218403  0.0  0.0  0.0  0.0  2.0\n",
      "neutral            3099.0  0.169732  0.401225  0.0  0.0  0.0  0.0  3.0\n",
      "positive           2363.0  0.098603  0.303818  0.0  0.0  0.0  0.0  2.0\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/matplotlib-2.2.2+772.g640ecd9-py3.5-linux-x86_64.egg/matplotlib/axes/_axes.py:6178: UserWarning: The 'normed' kwarg is deprecated, and has been replaced by the 'density' kwarg.\n",
      "  warnings.warn(\"The 'normed' kwarg is deprecated, and has been \"\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABCQAAAFUCAYAAAD8jO1vAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAAEZ0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yKzc3\nMi5nNjQwZWNkOSwgaHR0cDovL21hdHBsb3RsaWIub3JnL0J0boQAACAASURBVHic7d17WJR1/v/x\nF4mHdCAzqb6earMYD4BMVGCynjI1D6BdHbTUSl3TVitLxbR0N7fMpU3TytLMMq1Wt4Oo2UlbwUNc\naWYpYUkqJKWk1gCe8fP7wx+zjaAMMN5z8Pm4Lq7Luedzz3ze87nn7fjynpsQY4wRAAAAAACAhS7w\n9QQAAAAAAMD5h0ACAAAAAABYjkACAAAAAABYjkACAAAAAABYjkACAAAAAABYjkACAAAAAABYjkAi\nQHTu3Flz5sypcJzdbteKFSvOeDuYZWZmym63q6CgwNdTQTnOp2MRwYG+WzH6bmA6n45R+B96a8Xo\nreeWJ8fSe++9p+joaItmdH4jkAgQ//nPfzRgwIBK77d27Vp16dLlHMzIt6Kjo/Xee++5bXM4HFq7\ndq0uueQSH83K3ZAhQzR+/HhfT8NyZ6o7WI9FBC/6rjv6rm+sWLFCdrvd19MAvIbe6o7ear0/HksF\nBQWy2+3KzMx0G9OjRw99/vnnvpjeeSfU1xOAZxo0aHDW+48dO6ZatWqV2R4REXGupuR3atWqdV7V\nG2hYGwQa+m7F6Lv+40zHI+Bv6K0Vo7eeW568tnXq1FGdOnUsmA04Q8IPrFu3TgMHDtQNN9yguLg4\nDRgwQN98843bmNNPb7Pb7Vq4cKEeeughORwOTZo0qdzH/uMpST/99JPsdrs++ugjDRs2TG3atFGX\nLl20bNkyt3327dunMWPGKD4+XnFxcRo4cKC2bdvmUS3Hjx/XP/7xD/35z39WVFSU2rdvr3/84x+u\n+40xev3119W1a1dFR0ere/fuWrBggYwxbnN+++239eijj8rhcKhDhw567bXX3F6LY8eO6bHHHpPd\nbnf9z9Hpp7eVnmq1YcMG9erVSzExMRo4cKD27t2rL7/8Un369FFsbKzuuece7d27162O9PR03X77\n7YqJiVHHjh315JNPqqioyHX/wIEDNWnSJL3wwgu68cYbFR8frwkTJujw4cOSpPHjx2vt2rV6//33\nXXM8PXn1ltK6N2zYoP79+ysmJka9evXS+vXr3cbt3LlTI0aMUFxcnOLj4zV8+HDl5ua6jXn11VeV\nmJio2NhYPfjgg1q0aJHb/8zl5eVp5MiRSkxMVJs2bdS7d2998MEHrvvPVvcfj8V+/frpb3/7m9tz\nnzx5Uu3bt9err74qybNjBagq+i59tzq81XdnzZql7t27l/vYBQUFyszM1COPPCJJrppK/5dy4MCB\neuKJJ/Svf/1L7dq1cz3OsmXLdPvtt7uec9iwYdq5c+c5eR2A09Fb6a3VUVr3f//7X916662Kjo5W\n7969yzzfpk2b1K9fP0VHRys+Pl4TJ050q2n79u269957FRcXJ4fDod69e2vNmjWu+/94LCUmJkqS\nBg0aJLvdrs6dO0ty/8rGzp07ZbfbyxzLa9euVatWrVzrVJ3j7bxm4HOffPKJWbFihcnJyTHff/+9\nefzxx831119vDhw44BrTqVMn88orr7huR0ZGmhtuuMEsXLjQ7N692+zevdu1ffny5W7jSm/n5eWZ\nyMhIc9NNN5mVK1eaXbt2mX/+85+mdevWJi8vzxhjzOHDh023bt3M6NGjzdatW82PP/5opk2bZuLi\n4sy+ffsqrGXevHmmffv2JjMz0+zZs8ds3rzZvPPOO677Z8yYYTp37mxWrVplcnNzzSeffGLi4+PN\nokWL3Obcrl07s3jxYrNr1y7z+uuvm8jISLNx40ZjjDH79+83UVFR5vXXXzf79u1zzeuLL74wkZGR\nrtvvvvuuadGihRkwYID5+uuvzdatW03Xrl1N//79zYABA8xXX31lvvvuO3PLLbeYUaNGuZ5/3bp1\npk2bNuatt94yu3btMps2bTK33nqreeihh1xjBgwYYOLi4swzzzxjcnJyzJo1a0xcXJx56aWXjDHG\nOJ1OM3DgQPPQQw+55nj06NFyX7OlS5ea2NjYs/7Mnj37jK95ad3JyckmIyPD7Ny504wdO9Zcf/31\nprCw0BhjzL59+0xCQoKZMmWK2b59u/nhhx/M2LFjTadOnczhw4eNMcZ8+OGHpnXr1mbhwoXmxx9/\nNAsXLjQ33HCDiYyMdD1Xdna2efPNN813331ndu/ebRYuXGhatmxpNmzYUGHdfzwW33rrLXPDDTeY\nY8eOuR57/fr1pkWLFuaXX37x+FgBqoq+S9/1h747c+ZM061bt3Ifu3T+b7/9tuv2vn37jNPpdL0e\nsbGx5u9//7vZsWOH2b59uzHGmP/85z9m1apVZvfu3Wbbtm1mxIgR5uabb3Z7LU4/ZgFvobfSW73R\nW7t27Wr++9//mh9++MGMHz/exMbGml9//dUYY8zevXtNbGysSUlJMd9//71Zt26d6dy5s3nwwQdd\nj9OrVy/z6KOPmh07dpjdu3ebzz77zGRmZrqtS+mxlJ2dbSIjI83HH39s9u3bZ/bv3+96zaOiolz7\n3H777WbKlClu8x0zZowZMmSIMab6x9v5jEDCD508edLEx8ebpUuXuraV17yfeOKJMvt60rzfeOMN\n1/3Hjx83MTEx5r333jPGnPog06lTJ3PixAm3x+3Ro4eZO3duhXOfMmWKGTRokDl58mSZ+w4dOmRi\nYmLM+vXr3bbPnTvX9OjRw23OTz/9tNuYm2++2cyaNct1Oyoqyrz77rtuY8pr3pGRka4PacYYM3/+\nfBMZGWm+/fZb17YFCxaY6667znX77rvvNtOnT3d77C1btpjIyEhXkxowYIDp06eP25iJEyeaAQMG\nuG4PHjzYpKSklHkdTldYWGh27dp11p+DBw+ecf/Suj/77DPXtl9++cVERka6mu/zzz9v+vfv77bf\n8ePHTVxcnPnwww+NMcbceeedZvz48W5jUlJS3AKJ8owcOdJMnDixwrr/eCz+9ttvpnXr1ubTTz91\ne6777rvPGOP5sQJ4C32XvuuLvltRIGGMMcuXLy+3Dw8YMMB079693HU/vVa73e76B5AxBBKwDr2V\n3lqV3lq6hsYYc+zYMfPnP//ZvPjii8YYY5577jnTqVMnc/z4cdeY9PR0ExkZ6QqzHA6H22Oc7o/H\n0r59+0xkZKT54osv3MacHkgsXLjQJCQkuJ63uLjYxMbGmrS0NGNM9Y+38xnXkPADeXl5mjlzpr7+\n+mvt379fxhgdPnxY+fn5Z92vqld+bdmypevPoaGhatCggX799VdJ0tatW/XLL7/ouuuuc9vn6NGj\nZU7vL0/fvn01ePBgde3aVe3atVP79u3VoUMH1ahRQz/88IOOHDmiESNGKCQkxLXPiRMnypyG/8c5\nStKll17qmmNlhIaG6uqrr3bdbtiwoSS5fQ2hYcOGcjqdru8sbt26VVu2bNEbb7zhGlM6v927d7u+\n+9iiRYsyc/zqq68qPUebzSabzVbp/U73x9fs0ksvlSS3dd2yZYscDofbPocPH3ata05Ojvr27et2\nf2xsrN5//3238S+++KI+//xzFRQU6Pjx4zp27Jji4+MrNdeLLrpInTp1Ulpamrp06aLDhw/r448/\ndn2NozLHClAV9F36rj/03eqKiopyW1dJ+u677/TCCy/ou+++08GDByWdei3z8/MVFxfnlecFzoTe\nSm/1Rm+NjY11/blmzZqKjo5WTk6OpFOfV2NjYxUa+r9/xl5//fWSpB07dqhZs2a677779Pjjj+v9\n999XfHy8unXr5vbaVUWPHj00depUZWRkqFOnTvrkk08UEhKim2++WVL1j7fzGYGEHxg+fLgaNGig\nyZMn6/LLL1fNmjU1YMAAHT9+/Kz71a1bt0rPV7NmTbfbISEhruZ08uRJXXPNNZo5c2aZ/TxpMK1b\nt9aqVau0bt06bdiwQRMmTNBVV13l9p26F154QU2bNq30HE+ePFnh85/uggsu0AUX/O9SKaV/afzx\n8Uu3/fE1GD58uHr16lXm8S677DKvzzEtLU2TJ08+65j7779fw4cPP+uYimpKTEzUhAkTyux30UUX\neTzXf/7zn/r88881btw4XXXVVbrwwguVmpqq33//3ePHKJWcnKzRo0fL6XS6vtdX2tQrc6wAVUHf\n9WyO9N1z23f/eByUOnHiRMUF/H8XXnih2+3Dhw9r8ODBuv766zV16lQ1bNhQISEh6tmzZ4XHNuAN\n9FbP5khvPXtvra5Ro0YpKSlJ6enpWrdunV566SVNmDBBd999d5Uf8+KLL1aHDh20dOlS13+qdevW\nzXXhy+oeb+czAgkfO3jwoHbs2KHXXntN7dq1kyTt3bu3SsmpN0RFRWnZsmUKDw/XxRdfXKXHsNls\n6tatm7p166Y777xTffr00ffff6+rr75atWvX1p49e1wXkKmqmjVrqqSkpFqPcSatW7fWjz/+qCuu\nuKJaj+PpHDt37qw2bdqcdUxlQoPytG7dWsuWLdP//d//nfEq7M2bN9eWLVt05513urZt2bLFbczG\njRuVnJysHj16SDrVfHfu3Ol2xWxP6+7QoYPq1q2rlStX6tNPP1XXrl1dH0i8eawAp6PvVg19t3I8\n6buXXHKJ9u/fr5MnT7r+oXH6BdBK/6FQUlKiGjVqnPU5c3JydODAAT3yyCO68sorJZ3q41X5hwVQ\nWfTWqqG3lrVlyxb96U9/knTq4qLffvut6/Np8+bNtWzZMp04ccJ1lsSXX34pSW5nQVxxxRUaOHCg\nBg4cqKeeekr//ve/yw0kSnusJ30yOTlZY8aMUU5OjjZs2KD58+e77vPG8Xa+IpDwsYsuukgNGjTQ\nv//9bzVq1Ei//fabUlNTffZrZnr37q358+drxIgReuihh9S0aVMVFBQoIyNDN954Y5nTkE43b948\nXXbZZWrZsqVq1qyptLQ0XXjhhWrUqJHq1aunYcOGKTU1VcYYJSQk6OjRo8rKytK+fft0//33ezzP\nJk2aKDMzU+3bt1fNmjUr/BVSlfHggw9q6NChuvzyy9W7d2/VqVNHP/74oz7//HNNmTKlUnNcv369\ncnNzZbPZFBYWViaBlrx3etvZDBw4UO+++65GjRql+++/XxEREfr555+1atUq9e/fX1deeaXuuece\njR07VtHR0Wrbtq3Wr1+v1atXuz3On/70J3322We66aabVLduXc2fP1+//PKL2+vvad01a9ZUz549\ntXDhQuXk5Lh+u4Ykrx4rwOnou/Rdf+m78fHxKi4u1syZM3Xrrbdq69ateuedd8rUJEmrV69WXFyc\nateurXr16pX7nI0aNVKtWrX05ptv6t5779VPP/2k1NRUt/9VBc4Veiu91Vu99eWXX9bFF1+sJk2a\naN68eXI6na5A4u6779aCBQv0xBNPaMiQISooKNDf/vY3de/eXc2aNVNxcbGee+45de3aVY0bN9aB\nAwe0ceNGNW/evNznql+/vsLCwrR27VpdffXVqlWr1hlDk44dO+rCCy/UI488ossuu8ztK8vVPd7O\nZ/wN5WMXXHCBnn/+ee3atUtJSUkaP368BgwY4HYalZXq1KmjhQsX6pprrtHYsWPVvXt3jR49Wrt2\n7XJ9P/Zs6tatq3nz5um2225TcnKyNm/erJdffln169eXJI0cOVJjxozRokWL1KtXLw0aNEhLlixx\nfeDyVEpKirKysnTTTTepbdu2Var1TNq2bavXXntN33zzjfr166e+fftq+vTplf590Pfee68aNGig\n5ORktW3btkrfxfOWhg0b6u2331bdunU1YsQI9ejRQxMmTFBRUZHCw8MlSbfccosefvhhvfjii0pO\nTta6dev0l7/8RbVr13Y9zmOPPabLL79cgwYN0j333KOGDRuqZ8+ebs9VmbqTk5P1/fffq2HDhkpI\nSHC7z1vHCnA6+i591wqe9N2rrrpKU6ZMUVpamuvXKI8ePdrtcVq1aqX77rtPkydP1o033njWf0Q0\naNBAqampysjIUI8ePfTMM89o3Lhxbt+1Bs4Veiu91VvGjRunGTNmKDk5Wd98841mz57tumbGpZde\nqldffVU7d+5U37599fDDDyshIUFPPfWUpFPX2vjtt9/02GOPqXv37hoxYoRatGhx1q+STJo0SR99\n9JE6duxY5npqf1SrVi11795d2dnZSkpKcrt+SHWPt/NZiOEKcQDO4O9//7s2btxY5vd6AwAAAN6U\nmZmpQYMGae3atZUOTRC4iMwBSDr1Hb358+erQ4cOql27tjIyMrRkyRKNGzfO11MDAAAAEIQ8CiRW\nrFihRYsWKTs7W0eOHFFWVtZZx6enp2vatGnKy8tTs2bNNH78eC5MFySGDh2qTZs2lXtf79699eST\nT1o8I3hLSEiIMjMzNW/ePB0+fNj13q3OFYkBVB99FwC8j94K+AePvrKRkZGh33//XUeOHNGkSZPO\nGkjk5eWpV69eevLJJ3XLLbfoo48+0qRJk7R8+XK++x0E9u7dqyNHjpR7n81m0yWXXGLxjAAguNF3\nAcD76K2Af6jUNSQyMzN13333nTWQmDlzpr744gu99dZbrm133XWXbrzxRo0cObJ6swUAAAAAAEHB\n679lIzs7W61bt3bb1qpVK2VnZ1e474kT5+Z38AIA6LEAcC7RYwGg8rx+Ucvi4mKFhYW5bQsPD9eO\nHTsq3PfgwUNVft6IiDAVFBRWeX9/EAw1SNThb4KhjmCoQapeHRERYRUPqsD53mOl4KgjGGqQqMPf\nBEMd9Fjfow7/EQw1SNThb85Fn/X6GRL16tVTYaH7JJ1Op2w2m7efCgAAAAAABCivBxItWrQoc42J\nrKwstWjRwttPBQAAAAAAApRHgURJSYmOHj2q48ePS5KOHj2qo0ePqrzrYfbp00dbt27V8uXLdezY\nMaWlpSkrK0t9+vTx7swBAAAAAEDA8iiQWLp0qWJiYjRkyBCVlJQoJiZGMTEx2rNnj9LS0uRwOFxj\nmzVrplmzZmn27Nm67rrr9Morr+iFF17gV34CAAAAAAAXjy5qeeutt+rWW28t974mTZooKSnJbVv7\n9u3Vvn376s8OAAAAAAAEJa9fQwIAAAAAAKAiBBIAAAAAAMByBBIAAAAAAMByBBIAAAAAAMByBBIA\nAAAAAMByBBIAAAAAAMByHv3az0Cw5L8bVFR4xNfTqBZbWJ2Ar0GS7uvdyddTAAAAAAD4Oc6QAAAA\nAAAAliOQAAAAAAAAliOQAAAAAAAAliOQAAAAAAAAliOQAAAAAAAAliOQAAAAAAAAliOQAAAAAAAA\nliOQAAAAAAAAliOQAAAAAAAAliOQAAAAAAAAliOQAAAAAAAAliOQAAAAAAAAlgv19QQAAACA89m6\nnL0qKjzi62lUm815KCjq6BYR5uspAOcNzpAAAAAAAACWI5AAAAAAAACWI5AAAAAAAACWI5AAAAAA\nAACWI5AAAAAAAACWI5AAAAAAAACWI5AAAAAAAACWI5AAAAAAAACWI5AAAAAAAACWI5AAAAAAAACW\nI5AAAAAAAACWI5AAAAAAAACWI5AAAAAAAACWI5AAAAAAAACWI5AAAAAAAACWI5AAAAAAAACWI5AA\nAAAAAACWC/X1BAAA/u+dLTtVVHTE19OoNlt+nYCvY2hES19PAQAAwCs4QwIAAAAAAFiOQAIAAAAA\nAFiOQAIAAAAAAFiOQAIAAAAAAFiOQAIAAAAAAFiOQAIAAAAAAFjOo0CipKRE06ZNU0JCghwOh0aN\nGqUDBw6ccfy8efPUpUsXORwOde3aVYsWLfLahAEAAAAAQODzKJCYM2eOVq9erSVLlig9PV2SNG7c\nuHLHrlq1SrNmzdKzzz6rzZs3a9q0aUpNTdW6deu8N2sAAAAAABDQPAokFi9erKFDh6pp06YKCwvT\n2LFjlZGRoT179pQZm5ubK7vdrtjYWEmSw+GQ3W5Xdna2d2cOAAAAAAACVmhFA5xOp/Lz8xUVFeXa\n1qxZM9lsNmVnZ6tx48Zu43v06KF3331XmzZtksPh0FdffaVdu3bpz3/+c4WTufjiugoNrVGFMk6x\nhdWp8r7+IhhqkKSIiDBfT8ErqMN/BEMNkm/rqFaPzf9VNltw9KdgqIP3g3+hDv8RqD32e+ehoPkM\nGCx18H7wH9ThX7xdR4WBRHFxsSTJZrO5bQ8PD1dRUVGZ8Zdccom6deume+65RydPnpQkTZgwQZGR\nkRVO5uDBQx5N+kyKCo9Ua39fs4XVCfgaShUUFPp6CtUWERFGHX4iGGqQqleHN5p/tXtsUeD3J5ut\nTlDUcb6/H/wJdfiPgO+xQfAZkM+y/iMY3tMSdfibc9FnKwwk6tWrJ0llwgen01kmpJCkl156ScuX\nL9cHH3yg5s2ba8eOHRoxYoRq166t22+/vSpzBwAAAAAAQabCa0iEh4erUaNG2rZtm2tbbm6uioqK\nZLfby4zftm2bunbtqquvvlohISG65ppr1KVLF33++efenTkAAAAAAAhYHl3U8o477tDcuXOVl5en\nwsJCpaamKjExUU2aNCkz9tprr9Wnn36qXbt2SZJycnL02WefqXXr1l6dOAAAAAAACFwVfmVDkoYN\nGyan06nbbrtNx44dU7t27ZSamipJSktL0+TJk7V582ZJ0pAhQ1RYWKjBgwfr4MGDuuiii9S9e3cN\nGzbs3FUBAAAAAAACikeBRI0aNZSSkqKUlJQy9yUlJSkpKel/DxgaqjFjxmjMmDHemyUAAAAAAAgq\nHn1lAwAAAAAAwJsIJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOUI\nJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOUIJAAA\nAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAA\ngOUIJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOUI\nJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOUIJAAA\nAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAA\ngOUIJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOU8\nCiRKSko0bdo0JSQkyOFwaNSoUTpw4MAZx+/fv18pKSmKj4/Xtddeq+TkZO3du9drkwYAAAAAAIHN\no0Bizpw5Wr16tZYsWaL09HRJ0rhx48ode/ToUd17772qWbOmVq5cqY0bN+rZZ59VvXr1vDdrAAAA\nAAAQ0DwKJBYvXqyhQ4eqadOmCgsL09ixY5WRkaE9e/aUGfv+++/L6XRq8uTJatCggS644AJdc801\nstlsXp88AAAAAAAITBUGEk6nU/n5+YqKinJta9asmWw2m7Kzs8uMz8zM1BVXXKHx48crPj5e3bt3\n1+uvv+7VSQMAAAAAgMAWWtGA4uJiSSpzhkN4eLiKiorKjD948KAyMzM1YcIETZ06Vdu3b9fQoUPV\noEEDJSUlnfW5Lr64rkJDa1Rm/m5sYXWqvK+/CIYaJCkiIszXU/AK6vAfwVCD5Ns6qtVj83+VzRYc\n/SkY6uD94F+ow38Eao/93nkoaD4DBksdvB/8B3X4F2/XUWEgUXrth9PDB6fTWe7XMOrVq6fLLrtM\n99xzjyQpOjpaSUlJWrVqVYWBxMGDhzyeeHmKCo9Ua39fs4XVCfgaShUUFPp6CtUWERFGHX4iGGqQ\nqleHN5p/tXtsUeD3J5utTlDUcb6/H/wJdfiPgO+xQfAZkM+y/iMY3tMSdfibc9FnK/zKRnh4uBo1\naqRt27a5tuXm5qqoqEh2u73M+JYtWyokJKTM9vK2AQAAAACA85NHF7W84447NHfuXOXl5amwsFCp\nqalKTExUkyZNyozt27evfvvtNy1atEglJSXKzs7WsmXL1LVrV69PHgAAAAAABCaPAolhw4apc+fO\nuu2229S+fXsZY5SamipJSktLk8PhcI1t3Lix5syZoyVLliguLk4PPvigRo0apR49epybCgAAAAAA\nQMCp8BoSklSjRg2lpKQoJSWlzH1JSUllrg0RHx+vDz74wDszBAAAAAAAQcejMyQAAAAAAAC8iUAC\nAAAAAABYjkACAAAAAABYjkACAAAAAABYjkACAAAAAABYjkACAAAAAABYjkACAAAAAABYjkACAAAA\nAABYjkACAAAAAABYjkACAAAAAABYjkACAAAAAABYjkACAAAAAABYjkACAAAAAABYjkACAAAAAABY\njkACAAAAAABYjkACAAAAAABYjkACAAAAAABYjkACAAAAAABYjkACAAAAAABYjkACAAAAAABYjkAC\nAAAAAABYjkACAAAAAABYjkACAAAAAABYjkACAAAAAABYjkACAAAAAABYjkACAAAAAABYjkACAAAA\nAABYjkACAAAAAABYjkACAAAAAABYjkACAAAAAABYjkACAAAAAABYjkACAAAAAABYjkACAAAAAABY\njkACAAAAAABYjkACAAAAAABYjkACAAAAAABYjkACAAAAAABYjkACAAAAAABYjkACAAAAAABYjkAC\nAAAAAABYjkACAAAAAABYjkACAAAAAABYjkACAAAAAABYjkACAAAAAABYjkACAAAAAABYzqNAoqSk\nRNOmTVNCQoIcDodGjRqlAwcOVLjfW2+9JbvdrpdeeqnaEwUAAAAAAMHDo0Bizpw5Wr16tZYsWaL0\n9HRJ0rhx4866z549ezR//nxFRkZWf5YAAAAAACCoeBRILF68WEOHDlXTpk0VFhamsWPHKiMjQ3v2\n7DnjPhMnTtTo0aNVv359r00WAAAAAAAEh9CKBjidTuXn5ysqKsq1rVmzZrLZbMrOzlbjxo3L7PPO\nO+/owgsvVI8ePfT22297PJmLL66r0NAaHo8/nS2szLlmWQAAFhJJREFUTpX39RfBUIMkRUSE+XoK\nXkEd/iMYapB8W0e1emz+r7LZgqM/BUMdvB/8C3X4j0Dtsd87DwXNZ8BgqYP3g/+gDv/i7ToqDCSK\ni4slSTabzW17eHi4ioqKyozPz8/X7NmztXjx4kpP5uDBQ5Xe54+KCo9Ua39fs4XVCfgaShUUFPp6\nCtUWERFGHX4iGGqQqleHN5p/tXtsUeD3J5utTlDUcb6/H/wJdfiPgO+xQfAZkM+y/iMY3tMSdfib\nc9FnK/zKRr169SSpTPjgdDrLhBSS9Pjjj2vEiBG67LLLqjJPAAAAAABwHqjwDInw8HA1atRI27Zt\nU8uWLSVJubm5Kioqkt1uLzN+3bp12rZtm6ZPny7pVJDx7bffau3atXrrrbe8PH0AAAAAABCIKgwk\nJOmOO+7Q3LlzFR8fr/r16ys1NVWJiYlq0qRJmbFr1qxxu/3QQw8pLi5OgwcP9s6MAQAAAABAwPPo\nt2wMGzZMnTt31m233ab27dvLGKPU1FRJUlpamhwOh2vs5Zdf7vZTq1Yt2Ww2NWzY8NxUAAAAAAAA\nAo5HZ0jUqFFDKSkpSklJKXNfUlKSkpKSzrjvm2++WfXZAQAAAACAoOTRGRIAAAAAAADeRCABAAAA\nAAAsRyABAAAAAAAsRyABAAAAAAAsRyABAAAAAAAsRyABAAAAAAAsRyABAAAAAAAsRyABAAAAAAAs\nRyABAAAAAAAsRyABAAAAAAAsRyABAAAAAAAsRyABAAAAAAAsRyABAAAAAAAsRyABAAAAAAAsRyAB\nAAAAAAAsRyABAAAAAAAsRyABAAAAAAAsRyABAAAAAAAsRyABAAAAAAAsRyABAAAAAAAsRyABAAAA\nAAAsRyABAAAAAAAsRyABAAAAAAAsRyABAAAAAAAsRyABAAAAAAAsRyABAAAAAAAsRyABAAAAAAAs\nRyABAAAAAAAsRyABAAAAAAAsRyABAAAAAAAsRyABAAAAAAAsRyABAAAAAAAsRyABAAAAAAAsRyAB\nAAAAAAAsRyABAAAAAAAsRyABAAAAAAAsRyABAAAAAAAsRyABAAAAAAAsRyABAAAAAAAsRyABAAAA\nAAAsRyABAAAAAAAsRyABAAAAAAAsRyABAAAAAAAsRyABAAAAAAAs51EgUVJSomnTpikhIUEOh0Oj\nRo3SgQMHyh27Zs0aDRo0SPHx8br++ut11113aePGjV6dNAAAAAAACGweBRJz5szR6tWrtWTJEqWn\np0uSxo0bV+7Y33//XQMHDtSnn36qDRs2qFevXvrLX/6in3/+2XuzBgAAAAAAAc2jQGLx4sUaOnSo\nmjZtqrCwMI0dO1YZGRnas2dPmbFJSUm6+eabFR4ertDQUN11112qW7euvv32W69PHgAAAAAABKbQ\nigY4nU7l5+crKirKta1Zs2ay2WzKzs5W48aNz7r/9u3bdfDgQUVGRlY4mYsvrqvQ0BoeTLt8trA6\nVd7XXwRDDZIUERHm6yl4BXX4j2CoQfJtHdXqsfm/ymYLjv4UDHXwfvAv1OE/ArXHfu88FDSfAYOl\nDt4P/oM6/Iu366gwkCguLpYk2Ww2t+3h4eEqKio667779+/Xgw8+qMGDB+vKK6+scDIHDx6qcMzZ\nFBUeqdb+vmYLqxPwNZQqKCj09RSqLSIijDr8RDDUIFWvDm80/2r32KLA7082W52gqON8fz/4E+rw\nHwHfY4PgMyCfZf1HMLynJerwN+eiz1b4lY169epJUpnwwel0lgkp/mjv3r0aNGiQ2rVrp0cffbQy\ncwUAAAAAAEGuwkAiPDxcjRo10rZt21zbcnNzVVRUJLvdXu4+P/30k+6++261b99ekyZNUkhIiPdm\nDAAAAAAAAp5HF7W84447NHfuXOXl5amwsFCpqalKTExUkyZNyozNycnRXXfdpZ49eyolJcXrEwYA\nAAAAAIHPo0Bi2LBh6ty5s2677Ta1b99exhilpqZKktLS0uRwOFxjX331Ve3du1cLFiyQw+Fw/aSl\npZ2bCgAAAAAAQMCp8KKWklSjRg2lpKSUe8ZDUlKSkpKSXLenTp2qqVOnem+GAAAAAAAg6Hh0hgQA\nAAAAAIA3EUgAAAAAAADLEUgAAAAAAADLEUgAAAAAAADLEUgAAAAAAADLEUgAAAAAAADLEUgAAAAA\nAADLhfp6AgAAwHPvbNmpoqIjvp5Gtdny6wRFHUMjWvp6CgAABCzOkAAAAAAAAJYjkAAAAAAAAJYj\nkAAAAAAAAJYjkAAAAAAAAJYjkAAAAAAAAJYjkAAAAAAAAJYjkAAAAAAAAJYjkAAAAAAAAJYjkAAA\nAAAAAJYjkAAAAAAAAJYjkAAAAAAAAJYjkAAAAAAAAJYjkAAAAAAAAJYjkAAAAAAAAJYL9fUEAAAA\nAMBfrMvZq6LCI76eRrXYnIcCvgZJ6hYR5usp4BzjDAkAAAAAAGA5AgkAAAAAAGA5AgkAAAAAAGA5\nAgkAAAAAAGA5AgkAAAAAAGA5AgkAAAAAAGA5AgkAAAAAAGA5AgkAAAAAAGA5AgkAAAAAAGA5AgkA\nAAAAAGA5AgkAAAAAAGA5AgkAAAAAAGC5UF9PAMFnyX83qKjwiK+nUW22sDpBUcd9vTv5egoAAAAA\nUAZnSAAAAAAAAMsRSAAAAAAAAMsRSAAAAAAAAMsRSAAAAAAAAMsRSAAAAAAAAMsRSAAAAAAAAMsR\nSAAAAAAAAMsRSAAAAAAAAMt5FEiUlJRo2rRpSkhIkMPh0KhRo3TgwIEzjk9PT1fPnj0VExOjXr16\nae3atV6bMAAAAAAACHyhngyaM2eOVq9erSVLlqh+/fqaMGGCxo0bp1dffbXM2Ly8PI0aNUpPPvmk\nbrnlFn300UcaOXKkli9friZNmni9AAAAAF95Z8tOFRUd8fU0qs2WXyfg6xga0dLXUwAAVJJHZ0gs\nXrxYQ4cOVdOmTRUWFqaxY8cqIyNDe/bsKTP2/fffV+vWrZWcnKxatWopKSlJrVq10gcffOD1yQMA\nAAAAgMBU4RkSTqdT+fn5ioqKcm1r1qyZbDabsrOz1bhxY7fx2dnZat26tdu2Vq1aKTs720tTBgAA\nAAAEu3U5e1VUGNhnb0mSzXkoKOroFhHm9cesMJAoLi6WJNlsNrft4eHhKioqKnd8WFhYmbE7duyo\ncDIR1Sjw9o5tq7wvEOyq897yF8FQg+TbOqrz3P2C5PUHENwCtccGy99xAIKft/tVhV/ZqFevniSV\nCR+cTmeZkKJ0fGFhoUdjAQAAAADA+anCQCI8PFyNGjXStm3bXNtyc3NVVFQku91eZnyLFi2UlZXl\nti0rK0stWrTwwnQBAAAAAEAw8OiilnfccYfmzp2rvLw8FRYWKjU1VYmJieX+1ow+ffpo69atWr58\nuY4dO6a0tDRlZWWpT58+Xp88AAAAAAAITCHGGFPRoJKSEj377LN67733dOzYMbVr105PPvmkGjRo\noLS0NE2ePFmbN292jU9PT9e0adOUl5enpk2b6rHHHlNiYuI5LQQAAAAAAAQOjwIJAAAAAAAAb/Lo\nKxsAAAAAAADeRCABAAAAAAAsFxCBRElJiaZNm6aEhAQ5HA6NGjVKBw4cOOP49PR09ezZUzExMerV\nq5fWrl1r4WzPrDJ1ZGZmym63y+FwuH769etn8YzLWrFihe666y5de+21atWqVYXj/XUtKlOHv65F\namqqevbsqWuvvVaJiYl6/PHH9dtvv511H39cj8rW4a/rIUnTp09X586dde2116pt27Z68MEHlZ+f\nf8bx/rIe9Fj/Oo6Coc8GQ4+VgqPP0mP9Yy2Coc/SY/1nLeix/rUewdJnfdZjTQB46aWXTNeuXU1u\nbq5xOp1m5MiRZsiQIeWOzc3NNTExMeaDDz4wR48eNUuXLjVt2rQxeXl5Fs+6rMrU8cUXX5iWLVta\nPMOKpaenm2XLlpklS5ZUOD9/XovK1OGva/Gvf/3LbNu2zRw7dszs37/fDBkyxNx///1nHO+v61HZ\nOvx1PYwxZseOHcbpdBpjjDl06JB5+umnzZ133lnuWH9aD3qsfwmGPhsMPdaY4Oiz9Fj/WItg6LP0\nWP9ZC3qsf61HsPRZX/XYgAgkOnbsaBYvXuy6vXv3bhMZGWl++umnMmOff/55079/f7dt/fv3N7Nm\nzTrn86xIZerw1wO1lCfz8+e1KOVJHf6+FqXWrFljHA7HGe8PhPUwpuI6AmU9iouLzTPPPGNuuOGG\ncu/3p/Wgx/qnYOizwdRjjQmOPkuP9c1aBEOfpcee4g9rUYoe61/rUSoY+qyVPdbvv7LhdDqVn5+v\nqKgo17ZmzZrJZrMpOzu7zPjs7Gy1bt3abVurVq3KHWulytYhnTo1rkOHDmrXrp2GDRvm8xoqy1/X\noioCYS02bNigFi1anPH+QFmPiuqQ/Hs9li1bpri4ODkcDi1YsEAjR44sd5y/rAc91j+PI0/563pU\nVqCsRTD0WXqs9bUEQ5+lx/6Pr9eiKgJlLYKhx0qB3Wd90WP9PpAoLi6WJNlsNrft4eHhKioqKnd8\nWFiYR2OtVNk6rrrqKi1dulSrVq3SypUrZbfbdc8992jv3r2WzNcb/HUtKisQ1uLjjz/WO++8o4kT\nJ55xTCCshyd1+Pt69O7dW5s2bdLatWs1cuRIRUZGljvOX9aDHuufx5Gn/HU9KiNQ1iIY+iw91jdr\nEQx9lh77P75ei8oKlLUIhh4rBX6f9UWP9ftAol69epJUpjin01mmKZaOLyws9GislSpbR0REhFq0\naKHQ0FCFh4fr0UcfVf369ZWenm7JfL3BX9eisvx9LVauXKknnnhCs2fPLpNU/pG/r4endfj7epSK\niIjQHXfcoeHDh5d7YSN/WQ96rH8fRxXx1/WojEBYi2Dos/RY361FMPRZeuz/+HotKisQ1iIYeqwU\nXH3Wyh7r94FEeHi4GjVqpG3btrm25ebmqqioSHa7vcz4Fi1aKCsry21bVlZWhafNnGuVrSMY+Ota\nBJN3331XkydP1uzZs5WQkHDWsf68HpWpI5CcOHFChw4d0r59+8rc5y/rQY8NbP66HsEkGPosPfYU\nX61FMPRZeuz/+Hotgk0w9FgpOPusZT22Sle5sFh5V/UdPHhwuWN3795tYmJizLJly/zq6qvGVK6O\n9evXm127dpmSkhJTVFRkZs6caeLi4kx+fr7Fs3Z34sQJc+TIEZORkWFatmxpjhw5Yo4cOWJOnjxZ\nZqw/r0Vl6vDXtXjjjTfMDTfcYLZs2eLReH9dj8rW4a/rUVJSYt58803z66+/GmOM+fnnn80DDzxg\nOnXqZI4fP15mvD+tBz3Wf44jY4KjzwZDjzUmOPosPdY/1iIY+iw91n/Wgh7rX+sRDH3Wlz02IAKJ\nEydOuK7yGRsba/7617+a/fv3G2OMWbp0qYmNjXUbv2bNGtOjRw8THR1tevToYTIyMnwx7TIqU8f8\n+fNNx44dTZs2bUxCQoIZPHiwxwf5ufTuu++ayMjIMj95eXkBtRaVqcNf1yIyMtK0atXKxMbGuv2U\nCpT1qGwd/roeJSUlZujQoSYhIcG0adPGJCYmmkceecTs3r3bGOPf60GP9Z/jyJjg6LPB0GONCY4+\nS4/1j7UIhj5Lj/WftaDH+td6BEOf9WWPDTHGmKqexgEAAAAAAFAVfn8NCQAAAAAAEHwIJAAAAAAA\ngOUIJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOUIJAAAAAAAgOUIJAAfy8zMlN1u\nV0FBga+nAgBBhx4LAOcOPRbVRSABnKagoEB2u12ZmZm+ngoABB16LACcO/RYBBoCCcCHjh8/7usp\nAEDQoscCwLlDj4U3EEjArxhjtGDBAt1yyy2KiopSu3btNH78eElSYWGhJkyYoPj4eEVHR6tfv37a\nvHmza98znTJmt9u1YsUKSdJPP/0ku92ujz76SMOGDVObNm3UpUsXLVu2zDU+MTFRkjRo0CDZ7XZ1\n7ty5wnm/9957io6OdttW+lxff/212/zWrFmjO++8U1FRUW7PW6qwsFDjxo1T27ZtFR0drc6dO2vO\nnDmevHwAcFb0WHosgHOHHkuPReWF+noCwB/NmDFDb775psaNG6eEhAQdOnRIGzZskCQ99thj2r59\nu2bMmKGGDRtq7ty5GjJkiD755BM1bNiwUs/z7LPPasyYMZo4caIWL16sxx57TA6HQ02aNFFaWpqS\nkpI0a9YsORwO1ahRw6s1Tps2TePGjVPz5s1Vu3Zt7dy50+3+GTNmaPv27XrllVd0ySWXaM+ePdq7\nd69X5wDg/ESPpccCOHfosfRYVB6BBPxGcXGx5s+frzFjxqhfv36u7a1atdKuXbv06aefav78+Wrb\ntq0k6amnntKXX36phQsX6uGHH67Ucw0aNEjdu3eXJI0ePVoLFy7Ul19+qSZNmqhBgwaSpIsuukgR\nERFequ5/HnjgAXXs2NF1+/RGvmfPHrVq1UoxMTGSpMaNG3t9DgDOP/TYU+ixAM4Feuwp9FhUFl/Z\ngN/IycnR0aNH1a5du3LvCwkJUVxcnGtbzZo1FRsbq5ycnEo/V8uWLV1/Dg0NVYMGDfTrr79WbeKV\nVNqgz6R///5auXKlevfuralTp7qSdQCoDnrsKfRYAOcCPfYUeiwqi0ACQeOCC04dzsYY17YzXWyn\nZs2abrdDQkLc9qvKc5++/4kTJ8ode+GFF571sTp06KDVq1dr6NChcjqdeuCBBzR69Ogqzw0AvIEe\nCwDnDj0W5ysCCfiN5s2bq1atWlq3bl2Z+66++moZY7Rp0ybXtuPHj+vrr7/W1VdfLUmuU9T27dvn\nGpOVlVXpeZQ2+ZMnT3q8T4MGDXT8+HEdPHiwWs/9x8dLTk7W1KlTlZqaqg8//FC///57lR8PAOix\n7o9HjwXgTfRY98ejx8JTBBLwG/Xq1dN9992nGTNm6J133tHu3buVlZWlefPm6YorrlDXrl01efJk\nbdiwQTt27NDEiRP1+++/6+6775YkXXHFFWrcuLFmzZqlnJwcbdy4UdOmTav0POrXr6+wsDCtXbtW\nBQUFHjXQmJgY1atXT88++6x2796tNWvWaPbs2ZV+bkmaPn26PvvsM+3atUs5OTn6+OOPFRERofDw\n8Co9HgBI9NhS9FgA5wI99hR6LCqLQAJ+ZfTo0frrX/+q1157TT179tSwYcO0Y8cOSdLTTz+t66+/\nXg8//LD69u2r3NxczZs3z3Vl4tDQUE2fPl0FBQXq27evpkyZokcffbRK85g0aZI++ugjdezYUX37\n9q1wfP369fXcc8/pq6++Uu/evfXyyy9rzJgxVXrumjVravr06erTp4/69eun/fv369VXX1VISEiV\nHg8AStFj6bEAzh16LD0WlRdiqvOFIwAAAAAAgCrgDAkAAAAAAGC5UF9PAAgEPXv2VH5+frn33X//\n/Ro+fLjFMwKA4EGPBYBzhx4Lf8ZXNgAP7Nmz54y//uiiiy5S/fr1LZ4RAAQPeiwAnDv0WPgzAgkA\nAAAAAGA5riEBAAAAAAAsRyABAAAAAAAsRyABAAAAAAAsRyABAAAAAAAs9/8AYxDHGw7gz/kAAAAA\nSUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 1080x360 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_dist(df_eda, 'count_urls')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Descriptive stats for count_emojis\n",
      "----------------------------------\n",
      "                    count      mean       std  min  25%  50%  75%   max\n",
      "airline_sentiment                                                      \n",
      "negative           9178.0  0.028547  0.258772  0.0  0.0  0.0  0.0  8.0 \n",
      "neutral            3099.0  0.075186  0.492514  0.0  0.0  0.0  0.0  13.0\n",
      "positive           2363.0  0.179856  1.256408  0.0  0.0  0.0  0.0  40.0\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/matplotlib-2.2.2+772.g640ecd9-py3.5-linux-x86_64.egg/matplotlib/axes/_axes.py:6178: UserWarning: The 'normed' kwarg is deprecated, and has been replaced by the 'density' kwarg.\n",
      "  warnings.warn(\"The 'normed' kwarg is deprecated, and has been \"\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABCQAAAFUCAYAAAD8jO1vAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nTkSuQmCC\n",
      "text/plain": [
       "<Figure size 1080x360 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_dist(df_eda, 'count_emojis')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "class CleanText(BaseEstimator, TransformerMixin):\n",
    "    def remove_mentions(self, input_text):\n",
    "        return re.sub(r'@\\w+', '', input_text)\n",
    "    \n",
    "    def remove_urls(self, input_text):\n",
    "        return re.sub(r'http.?://[^\\s]+[\\s]?', '', input_text)\n",
    "    \n",
    "    def emoji_oneword(self, input_text):\n",
    "        # By compressing the underscore, the emoji is kept as one word\n",
    "        return input_text.replace('_','')\n",
    "    \n",
    "    def remove_punctuation(self, input_text):\n",
    "        # Make translation table\n",
    "        punct = string.punctuation\n",
    "        trantab = str.maketrans(punct, len(punct)*' ')  # Every punctuation symbol will be replaced by a space\n",
    "        return input_text.translate(trantab)\n",
    "\n",
    "    def remove_digits(self, input_text):\n",
    "        return re.sub('\\d+', '', input_text)\n",
    "    \n",
    "    def to_lower(self, input_text):\n",
    "        return input_text.lower()\n",
    "    \n",
    "    def remove_stopwords(self, input_text):\n",
    "        stopwords_list = stopwords.words('english')\n",
    "        # Some words which might indicate a certain sentiment are kept via a whitelist\n",
    "        whitelist = [\"n't\", \"not\", \"no\"]\n",
    "        words = input_text.split() \n",
    "        clean_words = [word for word in words if (word not in stopwords_list or word in whitelist) and len(word) > 1] \n",
    "        return \" \".join(clean_words) \n",
    "    \n",
    "    def stemming(self, input_text):\n",
    "        porter = PorterStemmer()\n",
    "        words = input_text.split() \n",
    "        stemmed_words = [porter.stem(word) for word in words]\n",
    "        return \" \".join(stemmed_words)\n",
    "    \n",
    "    def fit(self, X, y=None, **fit_params):\n",
    "        return self\n",
    "    \n",
    "    def transform(self, X, **transform_params):\n",
    "        clean_X = X.apply(self.remove_mentions).apply(self.remove_urls).apply(self.emoji_oneword).apply(self.remove_punctuation).apply(self.remove_digits).apply(self.to_lower).apply(self.remove_stopwords).apply(self.stemming)\n",
    "        return clean_X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "22 records have no words left after text cleaning\n"
     ]
    }
   ],
   "source": [
    "from nltk.corpus import stopwords\n",
    "from nltk.stem import PorterStemmer\n",
    "from nltk.tokenize import word_tokenize\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "ct = CleanText()\n",
    "sr_clean = ct.fit_transform(df.text)\n",
    "sr_clean.sample(5)\n",
    "empty_clean = sr_clean == ''\n",
    "print('{} records have no words left after text cleaning'.format(sr_clean[empty_clean].count()))\n",
    "sr_clean.loc[empty_clean] = '[no_text]'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0 records have no words left after text cleaning\n",
      "0    flight stuff wrong noth worst flight ever                                                 \n",
      "1    yeah flight tomorrow go add min extra road snow                                           \n",
      "2    no worri custom servic get bad wrap spoke agent save huge amount time amp apolog yesterday\n",
      "3    miss one probabl aspen event fail                                                         \n",
      "4    thank ask                                                                                 \n",
      "Name: text, dtype: object\n"
     ]
    }
   ],
   "source": [
    "i=1\n",
    "empty_clean = sr_clean == ''\n",
    "\n",
    "\n",
    "print('{} records have no words left after text cleaning'.format(sr_clean[empty_clean].count()))\n",
    "sr_clean.loc[empty_clean] = '[no_text]'\n",
    "print (sr_clean[0:5])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['aa',\n",
       " 'aaaand',\n",
       " 'aaadvantag',\n",
       " 'aaalwaysl',\n",
       " 'aadavantag',\n",
       " 'aadelay',\n",
       " 'aadv',\n",
       " 'aadvantag',\n",
       " 'aafail',\n",
       " 'aal',\n",
       " 'aaron',\n",
       " 'aaso',\n",
       " 'aateam',\n",
       " 'ab',\n",
       " 'aback',\n",
       " 'abandon',\n",
       " 'abassinet',\n",
       " 'abbrev',\n",
       " 'abc',\n",
       " 'abcletjetbluestreamfe']"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cv = CountVectorizer()\n",
    "bow = cv.fit_transform(sr_clean)\n",
    "#printing only first 20\n",
    "cv.get_feature_names()[0:20]\n",
    "#bow.sum(axis=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "      word  freq\n",
      "0   flight  4879\n",
      "1   thank   1703\n",
      "2   get     1623\n",
      "3   not     1586\n",
      "4   no      1507\n",
      "5   hour    1167\n",
      "6   cancel  1065\n",
      "7   help    1062\n",
      "8   delay   1010\n",
      "9   servic  1002\n",
      "10  time    987 \n",
      "11  custom  947 \n",
      "12  call    792 \n",
      "13  bag     771 \n",
      "14  wait    754 \n",
      "15  plane   735 \n",
      "16  us      705 \n",
      "17  fli     702 \n",
      "18  need    688 \n",
      "19  hold    687 \n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAuYAAAJUCAYAAABZm+MtAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<Figure size 864x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "word_freq = dict(zip(cv.get_feature_names(), np.asarray(bow.sum(axis=0)).ravel()))\n",
    "word_counter = collections.Counter(word_freq)\n",
    "word_counter_df = pd.DataFrame(word_counter.most_common(20), columns = ['word', 'freq'])\n",
    "print(word_counter_df)\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(12, 10))\n",
    "sns.barplot(x=\"word\", y=\"freq\", data=word_counter_df, palette=\"PuBuGn_d\", ax=ax)\n",
    "plt.show();\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['count_capital_words',\n",
       " 'count_emojis',\n",
       " 'count_excl_quest_marks',\n",
       " 'count_hashtags',\n",
       " 'count_mentions',\n",
       " 'count_urls',\n",
       " 'count_words',\n",
       " 'airline_sentiment',\n",
       " 'clean_text']"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_model = df_eda\n",
    "df_model['clean_text'] = sr_clean\n",
    "df_model.columns.tolist()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "class ColumnExtractor(TransformerMixin, BaseEstimator):\n",
    "    def __init__(self, cols):\n",
    "        self.cols = cols\n",
    "\n",
    "    def transform(self, X, **transform_params):\n",
    "        #print(\"I am in\")\n",
    "        \n",
    "        #print(X[self.cols].dtype())\n",
    "        return X[self.cols]\n",
    "\n",
    "    def fit(self, X, y=None, **fit_params):\n",
    "        return self"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(df_model.drop('airline_sentiment', axis=1), df_model.airline_sentiment, test_size=0.1, random_state=37)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def grid_vect(clf, parameters_clf, X_train, X_test, parameters_text=None, vect=None, is_w2v=False):\n",
    "    \n",
    "    textcountscols = ['count_capital_words','count_emojis','count_excl_quest_marks','count_hashtags'\n",
    "                      ,'count_mentions','count_urls','count_words']\n",
    "    \n",
    "    if is_w2v:\n",
    "        w2vcols = []\n",
    "        print(\"not good\")\n",
    "        for i in range(SIZE):\n",
    "            w2vcols.append(i)\n",
    "        features = FeatureUnion([('textcounts', ColumnExtractor(cols=textcountscols))\n",
    "                                 , ('w2v', ColumnExtractor(cols=w2vcols))]\n",
    "                                , n_jobs=-1)\n",
    "    else:\n",
    "        print(\"ok\")\n",
    "        features = FeatureUnion([('textcounts', ColumnExtractor(cols=textcountscols))\n",
    "                                 , ('pipe', Pipeline([('cleantext', ColumnExtractor(cols='clean_text')), ('vect', vect)]))]\n",
    "                                , n_jobs=1)\n",
    "\n",
    "    \n",
    "    pipeline = Pipeline([\n",
    "        ('features', features)\n",
    "        , ('clf', clf)\n",
    "    ])\n",
    "    \n",
    "    # Join the parameters dictionaries together\n",
    "    parameters = dict()\n",
    "    if parameters_text:\n",
    "        parameters.update(parameters_text)\n",
    "    parameters.update(parameters_clf)\n",
    "\n",
    "    # Make sure you have scikit-learn version 0.19 or higher to use multiple scoring metrics\n",
    "    grid_search = GridSearchCV(pipeline, parameters, n_jobs=-1, verbose=1, cv=5)\n",
    "    \n",
    "    print(\"Performing grid search...\")\n",
    "    print(\"pipeline:\", [name for name, _ in pipeline.steps])\n",
    "    print(\"parameters:\")\n",
    "    pprint(parameters)\n",
    "\n",
    "    t0 = time()\n",
    "    grid_search.fit(X_train, y_train)\n",
    "    print(\"done in %0.3fs\" % (time() - t0))\n",
    "    print()\n",
    "\n",
    "    print(\"Best CV score: %0.3f\" % grid_search.best_score_)\n",
    "    print(\"Best parameters set:\")\n",
    "    best_parameters = grid_search.best_estimator_.get_params()\n",
    "    for param_name in sorted(parameters.keys()):\n",
    "        print(\"\\t%s: %r\" % (param_name, best_parameters[param_name]))\n",
    "        \n",
    "    print(\"Test score with best_estimator_: %0.3f\" % grid_search.best_estimator_.score(X_test, y_test))\n",
    "    print(\"\\n\")\n",
    "    print(\"Test score with best_estimator_: %0.3f\" % grid_search.best_estimator_.score(X_train, y_train))\n",
    "    print(\"\\n\")\n",
    "    print(\"Classification Report Test Data\")\n",
    "    print(classification_report(y_test, grid_search.best_estimator_.predict(X_test)))\n",
    "    print(\"Classification Report Train Data\")\n",
    "    print(classification_report(y_train, grid_search.best_estimator_.predict(X_train)))                    \n",
    "    return grid_search"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Parameter grid settings for the vectorizers (Count and TFIDF)\n",
    "parameters_vect = {\n",
    "    'features__pipe__vect__max_df': (0.25, 0.5, 0.75),\n",
    "    'features__pipe__vect__ngram_range': ((1, 1), (1, 2)),\n",
    "    'features__pipe__vect__min_df': (1,2)\n",
    "}\n",
    "\n",
    "\n",
    "# Parameter grid settings for MultinomialNB\n",
    "parameters_mnb = {\n",
    "    'clf__alpha': (0.25, 0.5, 0.75)\n",
    "}\n",
    "\n",
    "\n",
    "# Parameter grid settings for LogisticRegression\n",
    "parameters_logreg = {\n",
    "    'clf__C': (1.05, 1.1, 1.15,1.2),\n",
    "    'clf__penalty': ('l1', 'l2')\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "mnb = MultinomialNB()\n",
    "logreg = LogisticRegression()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "countvect = CountVectorizer()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ok\n",
      "Performing grid search...\n",
      "pipeline: ['features', 'clf']\n",
      "parameters:\n",
      "{'clf__alpha': (0.25, 0.5, 0.75),\n",
      " 'features__pipe__vect__max_df': (0.25, 0.5, 0.75),\n",
      " 'features__pipe__vect__min_df': (1, 2),\n",
      " 'features__pipe__vect__ngram_range': ((1, 1), (1, 2))}\n",
      "Fitting 5 folds for each of 36 candidates, totalling 180 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done 180 out of 180 | elapsed:    5.8s finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "done in 6.774s\n",
      "\n",
      "Best CV score: 0.779\n",
      "Best parameters set:\n",
      "\tclf__alpha: 0.5\n",
      "\tfeatures__pipe__vect__max_df: 0.5\n",
      "\tfeatures__pipe__vect__min_df: 2\n",
      "\tfeatures__pipe__vect__ngram_range: (1, 2)\n",
      "Test score with best_estimator_: 0.787\n",
      "\n",
      "\n",
      "Test score with best_estimator_: 0.881\n",
      "\n",
      "\n",
      "Classification Report Test Data\n",
      "             precision    recall  f1-score   support\n",
      "\n",
      "   negative       0.85      0.90      0.87       941\n",
      "    neutral       0.58      0.50      0.54       287\n",
      "   positive       0.75      0.69      0.72       236\n",
      "\n",
      "avg / total       0.78      0.79      0.78      1464\n",
      "\n",
      "Classification Report Train Data\n",
      "             precision    recall  f1-score   support\n",
      "\n",
      "   negative       0.91      0.93      0.92      8237\n",
      "    neutral       0.80      0.74      0.77      2812\n",
      "   positive       0.85      0.87      0.86      2127\n",
      "\n",
      "avg / total       0.88      0.88      0.88     13176\n",
      "\n"
     ]
    }
   ],
   "source": [
    "best_mnb_countvect = grid_vect(mnb, parameters_mnb, X_train, X_test, parameters_text=parameters_vect, vect=countvect)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ok\n",
      "Performing grid search...\n",
      "pipeline: ['features', 'clf']\n",
      "parameters:\n",
      "{'clf__C': (1.05, 1.1, 1.15, 1.2),\n",
      " 'clf__penalty': ('l1', 'l2'),\n",
      " 'features__pipe__vect__max_df': (0.25, 0.5, 0.75),\n",
      " 'features__pipe__vect__min_df': (1, 2),\n",
      " 'features__pipe__vect__ngram_range': ((1, 1), (1, 2))}\n",
      "Fitting 5 folds for each of 96 candidates, totalling 480 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done 120 tasks      | elapsed:   32.4s\n",
      "[Parallel(n_jobs=-1)]: Done 370 tasks      | elapsed:  1.8min\n",
      "[Parallel(n_jobs=-1)]: Done 480 out of 480 | elapsed:  2.4min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "done in 145.882s\n",
      "\n",
      "Best CV score: 0.795\n",
      "Best parameters set:\n",
      "\tclf__C: 1.05\n",
      "\tclf__penalty: 'l2'\n",
      "\tfeatures__pipe__vect__max_df: 0.5\n",
      "\tfeatures__pipe__vect__min_df: 1\n",
      "\tfeatures__pipe__vect__ngram_range: (1, 2)\n",
      "Test score with best_estimator_: 0.811\n",
      "\n",
      "\n",
      "Test score with best_estimator_: 0.982\n",
      "\n",
      "\n",
      "Classification Report Test Data\n",
      "             precision    recall  f1-score   support\n",
      "\n",
      "   negative       0.87      0.91      0.89       941\n",
      "    neutral       0.64      0.59      0.61       287\n",
      "   positive       0.77      0.67      0.72       236\n",
      "\n",
      "avg / total       0.81      0.81      0.81      1464\n",
      "\n",
      "Classification Report Train Data\n",
      "             precision    recall  f1-score   support\n",
      "\n",
      "   negative       0.99      0.99      0.99      8237\n",
      "    neutral       0.96      0.96      0.96      2812\n",
      "   positive       0.97      0.97      0.97      2127\n",
      "\n",
      "avg / total       0.98      0.98      0.98     13176\n",
      "\n"
     ]
    }
   ],
   "source": [
    "best_logreg_countvect = grid_vect(logreg, parameters_logreg, X_train, X_test, parameters_text=parameters_vect, vect=countvect)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "TF-IDF (term frequency - inverse document frequency)  used to downweight these frequent words.\n",
    "\n",
    "tfidf(t,d)=tf(t,d)∗idf(t,d)\n",
    " \n",
    "where tf(t,d) is the number of times term t occurs in document d. The inverse document frequency can be computed as follows:\n",
    "\n",
    "idf(t,d)=lognd1+df(d,t)\n",
    " \n",
    "nd  is the number of documents\n",
    "df(d,t) is the number of documents that contain term t"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "tfidfvect = TfidfVectorizer()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ok\n",
      "Performing grid search...\n",
      "pipeline: ['features', 'clf']\n",
      "parameters:\n",
      "{'clf__alpha': (0.25, 0.5, 0.75),\n",
      " 'features__pipe__vect__max_df': (0.25, 0.5, 0.75),\n",
      " 'features__pipe__vect__min_df': (1, 2),\n",
      " 'features__pipe__vect__ngram_range': ((1, 1), (1, 2))}\n",
      "Fitting 5 folds for each of 36 candidates, totalling 180 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done 180 out of 180 | elapsed:    6.2s finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "done in 7.136s\n",
      "\n",
      "Best CV score: 0.748\n",
      "Best parameters set:\n",
      "\tclf__alpha: 0.25\n",
      "\tfeatures__pipe__vect__max_df: 0.25\n",
      "\tfeatures__pipe__vect__min_df: 2\n",
      "\tfeatures__pipe__vect__ngram_range: (1, 2)\n",
      "Test score with best_estimator_: 0.770\n",
      "\n",
      "\n",
      "Test score with best_estimator_: 0.848\n",
      "\n",
      "\n",
      "Classification Report Test Data\n",
      "             precision    recall  f1-score   support\n",
      "\n",
      "   negative       0.78      0.96      0.86       941\n",
      "    neutral       0.69      0.38      0.49       287\n",
      "   positive       0.78      0.47      0.58       236\n",
      "\n",
      "avg / total       0.76      0.77      0.74      1464\n",
      "\n",
      "Classification Report Train Data\n",
      "             precision    recall  f1-score   support\n",
      "\n",
      "   negative       0.84      0.97      0.90      8237\n",
      "    neutral       0.87      0.60      0.71      2812\n",
      "   positive       0.89      0.68      0.77      2127\n",
      "\n",
      "avg / total       0.85      0.85      0.84     13176\n",
      "\n"
     ]
    }
   ],
   "source": [
    "best_mnb_tfidf = grid_vect(mnb, parameters_mnb, X_train, X_test, parameters_text=parameters_vect, vect=tfidfvect)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ok\n",
      "Performing grid search...\n",
      "pipeline: ['features', 'clf']\n",
      "parameters:\n",
      "{'clf__C': (1.05, 1.1, 1.15, 1.2),\n",
      " 'clf__penalty': ('l1', 'l2'),\n",
      " 'features__pipe__vect__max_df': (0.25, 0.5, 0.75),\n",
      " 'features__pipe__vect__min_df': (1, 2),\n",
      " 'features__pipe__vect__ngram_range': ((1, 1), (1, 2))}\n",
      "Fitting 5 folds for each of 96 candidates, totalling 480 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done 120 tasks      | elapsed:   43.2s\n",
      "[Parallel(n_jobs=-1)]: Done 370 tasks      | elapsed:  2.4min\n",
      "[Parallel(n_jobs=-1)]: Done 480 out of 480 | elapsed:  3.2min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "done in 191.046s\n",
      "\n",
      "Best CV score: 0.790\n",
      "Best parameters set:\n",
      "\tclf__C: 1.15\n",
      "\tclf__penalty: 'l2'\n",
      "\tfeatures__pipe__vect__max_df: 0.5\n",
      "\tfeatures__pipe__vect__min_df: 2\n",
      "\tfeatures__pipe__vect__ngram_range: (1, 1)\n",
      "Test score with best_estimator_: 0.804\n",
      "\n",
      "\n",
      "Test score with best_estimator_: 0.843\n",
      "\n",
      "\n",
      "Classification Report Test Data\n",
      "             precision    recall  f1-score   support\n",
      "\n",
      "   negative       0.84      0.93      0.89       941\n",
      "    neutral       0.65      0.54      0.59       287\n",
      "   positive       0.79      0.61      0.69       236\n",
      "\n",
      "avg / total       0.80      0.80      0.80      1464\n",
      "\n",
      "Classification Report Train Data\n",
      "             precision    recall  f1-score   support\n",
      "\n",
      "   negative       0.85      0.96      0.90      8237\n",
      "    neutral       0.80      0.61      0.69      2812\n",
      "   positive       0.87      0.69      0.77      2127\n",
      "\n",
      "avg / total       0.84      0.84      0.84     13176\n",
      "\n"
     ]
    }
   ],
   "source": [
    "best_logreg_tfidf = grid_vect(logreg, parameters_logreg, X_train, X_test, parameters_text=parameters_vect, vect=tfidfvect)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "collapsed": false
   },
   "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>count_capital_words</th>\n",
       "      <th>count_emojis</th>\n",
       "      <th>count_excl_quest_marks</th>\n",
       "      <th>count_hashtags</th>\n",
       "      <th>count_mentions</th>\n",
       "      <th>count_urls</th>\n",
       "      <th>count_words</th>\n",
       "      <th>clean_text</th>\n",
       "      <th>clean_text_wordlist</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>9709</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>8</td>\n",
       "      <td>fli nashvil</td>\n",
       "      <td>[fli, nashvil]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12212</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>24</td>\n",
       "      <td>still think pj may sold love board us come</td>\n",
       "      <td>[still, think, pj, may, sold, love, board, us, come]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11438</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>25</td>\n",
       "      <td>appreci chang time min multipl time hour flight annoy seem like clue</td>\n",
       "      <td>[appreci, chang, time, min, multipl, time, hour, flight, annoy, seem, like, clue]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13844</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>24</td>\n",
       "      <td>unfortun patienc get boat leav pm hold watch flight fill</td>\n",
       "      <td>[unfortun, patienc, get, boat, leav, pm, hold, watch, flight, fill]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8781</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>30</td>\n",
       "      <td>flt lax fll ppl sleep id rather not disturb flight tri dl movi no go</td>\n",
       "      <td>[flt, lax, fll, ppl, sleep, id, rather, not, disturb, flight, tri, dl, movi, no, go]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       count_capital_words  count_emojis  count_excl_quest_marks  \\\n",
       "9709   0                    0             1                        \n",
       "12212  0                    0             1                        \n",
       "11438  0                    0             0                        \n",
       "13844  0                    0             0                        \n",
       "8781   1                    0             0                        \n",
       "\n",
       "       count_hashtags  count_mentions  count_urls  count_words  \\\n",
       "9709   1               1               0           8             \n",
       "12212  1               3               0           24            \n",
       "11438  0               1               0           25            \n",
       "13844  0               1               0           24            \n",
       "8781   0               1               1           30            \n",
       "\n",
       "                                                                 clean_text  \\\n",
       "9709   fli nashvil                                                            \n",
       "12212  still think pj may sold love board us come                             \n",
       "11438  appreci chang time min multipl time hour flight annoy seem like clue   \n",
       "13844  unfortun patienc get boat leav pm hold watch flight fill               \n",
       "8781   flt lax fll ppl sleep id rather not disturb flight tri dl movi no go   \n",
       "\n",
       "                                                                        clean_text_wordlist  \n",
       "9709   [fli, nashvil]                                                                        \n",
       "12212  [still, think, pj, may, sold, love, board, us, come]                                  \n",
       "11438  [appreci, chang, time, min, multipl, time, hour, flight, annoy, seem, like, clue]     \n",
       "13844  [unfortun, patienc, get, boat, leav, pm, hold, watch, flight, fill]                   \n",
       "8781   [flt, lax, fll, ppl, sleep, id, rather, not, disturb, flight, tri, dl, movi, no, go]  "
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "SIZE = 50\n",
    "\n",
    "X_train['clean_text_wordlist'] = X_train.clean_text.apply(lambda x : word_tokenize(x))\n",
    "X_test['clean_text_wordlist'] = X_test.clean_text.apply(lambda x : word_tokenize(x))\n",
    "X_train['clean_text_wordlist'].head()\n",
    "model = gensim.models.Word2Vec(X_train.clean_text_wordlist\n",
    "                 , min_count=1\n",
    "                 , size=SIZE\n",
    "                 , window=5\n",
    "                 , workers=4)\n",
    "\n",
    "X_train.head()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/ipykernel/__main__.py:1: DeprecationWarning: Call to deprecated `most_similar` (Method will be removed in 4.0.0, use self.wv.most_similar() instead).\n",
      "  if __name__ == '__main__':\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[('rebook', 0.9876677989959717),\n",
       " ('flightlat', 0.9841099977493286),\n",
       " ('tomorrow', 0.9824616312980652)]"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.most_similar('flight', topn=3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def compute_avg_w2v_vector(w2v_dict, tweet):\n",
    "    list_of_word_vectors = [w2v_dict[w] for w in tweet if w in w2v_dict.vocab.keys()]\n",
    "    \n",
    "    if len(list_of_word_vectors) == 0:\n",
    "        result = [0.0]*SIZE\n",
    "    else:\n",
    "        result = np.sum(list_of_word_vectors, axis=0) / len(list_of_word_vectors)\n",
    "        \n",
    "    return result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Int64Index([ 9709, 12212, 11438, 13844,  8781, 12474, 13046,  8006, 12949,\n",
      "            13903,\n",
      "            ...\n",
      "             6595,  5795, 11898,  4118, 12725,   988, 14587,  9036,  4459,\n",
      "             1935],\n",
      "           dtype='int64', length=13176)\n"
     ]
    }
   ],
   "source": [
    "X_train_w2v = X_train['clean_text_wordlist'].apply(lambda x: compute_avg_w2v_vector(model.wv, x))\n",
    "\n",
    "X_test_w2v = X_test['clean_text_wordlist'].apply(lambda x: compute_avg_w2v_vector(model.wv, x))\n",
    "print(X_train.index)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "collapsed": false
   },
   "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>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>...</th>\n",
       "      <th>47</th>\n",
       "      <th>48</th>\n",
       "      <th>49</th>\n",
       "      <th>count_capital_words</th>\n",
       "      <th>count_emojis</th>\n",
       "      <th>count_excl_quest_marks</th>\n",
       "      <th>count_hashtags</th>\n",
       "      <th>count_mentions</th>\n",
       "      <th>count_urls</th>\n",
       "      <th>count_words</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>9709</th>\n",
       "      <td>-0.233305</td>\n",
       "      <td>0.569635</td>\n",
       "      <td>-0.482216</td>\n",
       "      <td>0.324911</td>\n",
       "      <td>0.070882</td>\n",
       "      <td>0.329037</td>\n",
       "      <td>-0.027973</td>\n",
       "      <td>0.205157</td>\n",
       "      <td>0.599025</td>\n",
       "      <td>-0.177616</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.125286</td>\n",
       "      <td>-0.170067</td>\n",
       "      <td>0.699998</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12212</th>\n",
       "      <td>-0.222045</td>\n",
       "      <td>0.632078</td>\n",
       "      <td>-0.537508</td>\n",
       "      <td>0.366488</td>\n",
       "      <td>0.075658</td>\n",
       "      <td>0.287304</td>\n",
       "      <td>-0.001040</td>\n",
       "      <td>0.221889</td>\n",
       "      <td>0.575719</td>\n",
       "      <td>-0.242624</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.110047</td>\n",
       "      <td>-0.216184</td>\n",
       "      <td>0.651532</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11438</th>\n",
       "      <td>-0.180390</td>\n",
       "      <td>0.778435</td>\n",
       "      <td>-0.553215</td>\n",
       "      <td>0.410322</td>\n",
       "      <td>0.075281</td>\n",
       "      <td>0.247228</td>\n",
       "      <td>-0.023816</td>\n",
       "      <td>0.176980</td>\n",
       "      <td>0.609886</td>\n",
       "      <td>-0.347563</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.075041</td>\n",
       "      <td>-0.214839</td>\n",
       "      <td>0.613404</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>25</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13844</th>\n",
       "      <td>-0.151716</td>\n",
       "      <td>0.663203</td>\n",
       "      <td>-0.451885</td>\n",
       "      <td>0.341062</td>\n",
       "      <td>0.064204</td>\n",
       "      <td>0.197396</td>\n",
       "      <td>-0.022992</td>\n",
       "      <td>0.155782</td>\n",
       "      <td>0.516508</td>\n",
       "      <td>-0.298062</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.063865</td>\n",
       "      <td>-0.186452</td>\n",
       "      <td>0.476014</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8781</th>\n",
       "      <td>-0.174139</td>\n",
       "      <td>0.582780</td>\n",
       "      <td>-0.442664</td>\n",
       "      <td>0.308549</td>\n",
       "      <td>0.073706</td>\n",
       "      <td>0.225255</td>\n",
       "      <td>-0.025567</td>\n",
       "      <td>0.162181</td>\n",
       "      <td>0.512543</td>\n",
       "      <td>-0.240148</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.087707</td>\n",
       "      <td>-0.181182</td>\n",
       "      <td>0.521329</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 57 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "              0         1         2         3         4         5         6  \\\n",
       "9709  -0.233305  0.569635 -0.482216  0.324911  0.070882  0.329037 -0.027973   \n",
       "12212 -0.222045  0.632078 -0.537508  0.366488  0.075658  0.287304 -0.001040   \n",
       "11438 -0.180390  0.778435 -0.553215  0.410322  0.075281  0.247228 -0.023816   \n",
       "13844 -0.151716  0.663203 -0.451885  0.341062  0.064204  0.197396 -0.022992   \n",
       "8781  -0.174139  0.582780 -0.442664  0.308549  0.073706  0.225255 -0.025567   \n",
       "\n",
       "              7         8         9     ...             47        48  \\\n",
       "9709   0.205157  0.599025 -0.177616     ...      -0.125286 -0.170067   \n",
       "12212  0.221889  0.575719 -0.242624     ...      -0.110047 -0.216184   \n",
       "11438  0.176980  0.609886 -0.347563     ...      -0.075041 -0.214839   \n",
       "13844  0.155782  0.516508 -0.298062     ...      -0.063865 -0.186452   \n",
       "8781   0.162181  0.512543 -0.240148     ...      -0.087707 -0.181182   \n",
       "\n",
       "             49  count_capital_words  count_emojis  count_excl_quest_marks  \\\n",
       "9709   0.699998  0                    0             1                        \n",
       "12212  0.651532  0                    0             1                        \n",
       "11438  0.613404  0                    0             0                        \n",
       "13844  0.476014  0                    0             0                        \n",
       "8781   0.521329  1                    0             0                        \n",
       "\n",
       "       count_hashtags  count_mentions  count_urls  count_words  \n",
       "9709   1               1               0           8            \n",
       "12212  1               3               0           24           \n",
       "11438  0               1               0           25           \n",
       "13844  0               1               0           24           \n",
       "8781   0               1               1           30           \n",
       "\n",
       "[5 rows x 57 columns]"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_w2v = pd.DataFrame(X_train_w2v.values.tolist(), index= X_train.index)\n",
    "X_test_w2v = pd.DataFrame(X_test_w2v.values.tolist(), index= X_test.index)\n",
    "\n",
    "# Concatenate with the TextCounts variables\n",
    "X_train_w2v = pd.concat([X_train_w2v, X_train.drop(['clean_text', 'clean_text_wordlist'], axis=1)], axis=1)\n",
    "X_test_w2v = pd.concat([X_test_w2v, X_test.drop(['clean_text', 'clean_text_wordlist'], axis=1)], axis=1)\n",
    "X_train_w2v.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "not good\n",
      "Performing grid search...\n",
      "pipeline: ['features', 'clf']\n",
      "parameters:\n",
      "{'clf__C': (1.05, 1.1, 1.15, 1.2), 'clf__penalty': ('l1', 'l2')}\n",
      "Fitting 5 folds for each of 8 candidates, totalling 40 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "/home/15ucs099/anaconda3/lib/python3.5/site-packages/sklearn/externals/joblib/parallel.py:547: UserWarning: Multiprocessing-backed parallel loops cannot be nested, setting n_jobs=1\n",
      "  **self._backend_args)\n",
      "[Parallel(n_jobs=-1)]: Done   2 out of  40 | elapsed:    4.0s remaining:  1.3min\n",
      "[Parallel(n_jobs=-1)]: Done  40 out of  40 | elapsed:   34.6s finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "done in 62.532s\n",
      "\n",
      "Best CV score: 0.725\n",
      "Best parameters set:\n",
      "\tclf__C: 1.2\n",
      "\tclf__penalty: 'l1'\n",
      "Test score with best_estimator_: 0.726\n",
      "\n",
      "\n",
      "Test score with best_estimator_: 0.729\n",
      "\n",
      "\n",
      "Classification Report Test Data\n",
      "             precision    recall  f1-score   support\n",
      "\n",
      "   negative       0.76      0.93      0.84       941\n",
      "    neutral       0.52      0.36      0.42       287\n",
      "   positive       0.72      0.37      0.49       236\n",
      "\n",
      "avg / total       0.71      0.73      0.70      1464\n",
      "\n",
      "Classification Report Train Data\n",
      "             precision    recall  f1-score   support\n",
      "\n",
      "   negative       0.75      0.94      0.83      8237\n",
      "    neutral       0.61      0.37      0.46      2812\n",
      "   positive       0.73      0.39      0.51      2127\n",
      "\n",
      "avg / total       0.72      0.73      0.70     13176\n",
      "\n"
     ]
    }
   ],
   "source": [
    "best_logreg_w2v = grid_vect(logreg, parameters_logreg, X_train_w2v, X_test_w2v, is_w2v=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "textcountscols = ['count_capital_words','count_emojis','count_excl_quest_marks','count_hashtags'\n",
    "                      ,'count_mentions','count_urls','count_words']\n",
    "    \n",
    "features = FeatureUnion([('textcounts', ColumnExtractor(cols=textcountscols))\n",
    "                         , ('pipe', Pipeline([('cleantext', ColumnExtractor(cols='clean_text'))\n",
    "                                              , ('vect', CountVectorizer(max_df=0.5, min_df=1, ngram_range=(1,2)))]))]\n",
    "                       , n_jobs=-1)\n",
    "\n",
    "pipeline = Pipeline([\n",
    "    ('features', features)\n",
    "    , ('clf', LogisticRegression(C=1.0, penalty='l2'))\n",
    "])\n",
    "\n",
    "best_model = pipeline.fit(df_model.drop('airline_sentiment', axis=1), df_model.airline_sentiment)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['positive', 'positive', 'positive']"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "new_positive_tweets = pd.Series([\"Thank you @VirginAmerica for you amazing customer support team on Tuesday 11/28 at @EWRairport and returning my lost bag in less than 24h! #efficiencyiskey #virginamerica\"\n",
    "                      ,\"Love flying with you guys ask these years.  Sad that this will be the last trip 😂   @VirginAmerica  #LuxuryTravel\"\n",
    "                      ,\"Wow @VirginAmerica main cabin select is the way to fly!! This plane is nice and clean & I have tons of legroom! Wahoo! NYC bound! ✈️\"])\n",
    "\n",
    "df_counts_pos = tc.transform(new_positive_tweets)\n",
    "df_clean_pos = ct.transform(new_positive_tweets)\n",
    "df_model_pos = df_counts_pos\n",
    "df_model_pos['clean_text'] = df_clean_pos\n",
    "\n",
    "best_model.predict(df_model_pos).tolist()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['negative', 'negative', 'negative']"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "new_negative_tweets = pd.Series([\"@VirginAmerica shocked my initially with the service, but then went on to shock me further with no response to what my complaint was. #unacceptable @Delta @richardbranson\"\n",
    "                      ,\"@VirginAmerica this morning I was forced to repack a suitcase w a medical device because it was barely overweight - wasn't even given an option to pay extra. My spouses suitcase then burst at the seam with the added device and had to be taped shut. Awful experience so far!\"\n",
    "                      ,\"Board airplane home. Computer issue. Get off plane, traverse airport to gate on opp side. Get on new plane hour later. Plane too heavy. 8 volunteers get off plane. Ohhh the adventure of travel ✈️ @VirginAmerica\"])\n",
    "\n",
    "df_counts_neg = tc.transform(new_negative_tweets)\n",
    "df_clean_neg = ct.transform(new_negative_tweets)\n",
    "df_model_neg = df_counts_neg\n",
    "df_model_neg['clean_text'] = df_clean_neg\n",
    "\n",
    "best_model.predict(df_model_neg).tolist()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "collapsed": false
   },
   "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>count_capital_words</th>\n",
       "      <th>count_emojis</th>\n",
       "      <th>count_excl_quest_marks</th>\n",
       "      <th>count_hashtags</th>\n",
       "      <th>count_mentions</th>\n",
       "      <th>count_urls</th>\n",
       "      <th>count_words</th>\n",
       "      <th>clean_text</th>\n",
       "      <th>clean_text_wordlist</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>10620</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "      <td>chantilli pari franc</td>\n",
       "      <td>[chantilli, pari, franc]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13609</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>11</td>\n",
       "      <td>custom servic experi tweet qualiti assur</td>\n",
       "      <td>[custom, servic, experi, tweet, qualiti, assur]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2955</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>26</td>\n",
       "      <td>pl stop send text everi min say flight delay anoth min hr usual defend lame</td>\n",
       "      <td>[pl, stop, send, text, everi, min, say, flight, delay, anoth, min, hr, usual, defend, lame]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3585</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>9</td>\n",
       "      <td>absolut send</td>\n",
       "      <td>[absolut, send]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2087</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>come scoop</td>\n",
       "      <td>[come, scoop]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12382</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>22</td>\n",
       "      <td>help wrong stuck florida day need learn real airlin</td>\n",
       "      <td>[help, wrong, stuck, florida, day, need, learn, real, airlin]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4062</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>17</td>\n",
       "      <td>thank find lower fare alreadi carrier direct flight juli</td>\n",
       "      <td>[thank, find, lower, fare, alreadi, carrier, direct, flight, juli]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3567</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>26</td>\n",
       "      <td>check right hr mark board group kid age amp dnt want split sld</td>\n",
       "      <td>[check, right, hr, mark, board, group, kid, age, amp, dnt, want, split, sld]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10064</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>11</td>\n",
       "      <td>thank good back safe see soon</td>\n",
       "      <td>[thank, good, back, safe, see, soon]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1809</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "      <td>link flyfi</td>\n",
       "      <td>[link, flyfi]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14334</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>24</td>\n",
       "      <td>unrival access california flight uk airport nonstop one stop connect year round</td>\n",
       "      <td>[unrival, access, california, flight, uk, airport, nonstop, one, stop, connect, year, round]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10175</th>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>26</td>\n",
       "      <td>thank book spent extra hr ten no explan fli middl seat char instead memphi</td>\n",
       "      <td>[thank, book, spent, extra, hr, ten, no, explan, fli, middl, seat, char, instead, memphi]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3993</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>like</td>\n",
       "      <td>[like]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7851</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>16</td>\n",
       "      <td>thank help total appreci made super easi</td>\n",
       "      <td>[thank, help, total, appreci, made, super, easi]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1783</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>11</td>\n",
       "      <td>good grief flight cancel flightl hold sinc</td>\n",
       "      <td>[good, grief, flight, cancel, flightl, hold, sinc]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12791</th>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>20</td>\n",
       "      <td>airlin cancel flight oc fll flight today switch no notic no apolog done fli unitedairlin</td>\n",
       "      <td>[airlin, cancel, flight, oc, fll, flight, today, switch, no, notic, no, apolog, done, fli, unitedairlin]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14172</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>25</td>\n",
       "      <td>next time flight cancel flightl appreci indic email text carrier pigeon app even warn check</td>\n",
       "      <td>[next, time, flight, cancel, flightl, appreci, indic, email, text, carrier, pigeon, app, even, warn, check]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3878</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>16</td>\n",
       "      <td>tri check tomorrow flight websit keep time</td>\n",
       "      <td>[tri, check, tomorrow, flight, websit, keep, time]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12711</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>13</td>\n",
       "      <td>access dial make faster</td>\n",
       "      <td>[access, dial, make, faster]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11379</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>18</td>\n",
       "      <td>right good caught mechan failur bad no blanket ok</td>\n",
       "      <td>[right, good, caught, mechan, failur, bad, no, blanket, ok]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9061</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>16</td>\n",
       "      <td>no done yet someth onlin thx</td>\n",
       "      <td>[no, done, yet, someth, onlin, thx]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11739</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>16</td>\n",
       "      <td>saw one useless tsa agent inspect shoe</td>\n",
       "      <td>[saw, one, useless, tsa, agent, inspect, shoe]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13674</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>25</td>\n",
       "      <td>first good idea thatthi would happen not cancel flight earlier amateurish no plan</td>\n",
       "      <td>[first, good, idea, thatthi, would, happen, not, cancel, flight, earlier, amateurish, no, plan]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>533</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>thank</td>\n",
       "      <td>[thank]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9701</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>22</td>\n",
       "      <td>morn boston beauti san diego realli bum satellit broken</td>\n",
       "      <td>[morn, boston, beauti, san, diego, realli, bum, satellit, broken]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4603</th>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>19</td>\n",
       "      <td>departur sure cater strike jfk</td>\n",
       "      <td>[departur, sure, cater, strike, jfk]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5835</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>26</td>\n",
       "      <td>guess wait anoth hr hope rate goe amp reason finger cross cut close travel date</td>\n",
       "      <td>[guess, wait, anoth, hr, hope, rate, goe, amp, reason, finger, cross, cut, close, travel, date]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4027</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>26</td>\n",
       "      <td>book time call told confirm link even make chang rez</td>\n",
       "      <td>[book, time, call, told, confirm, link, even, make, chang, rez]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10554</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>16</td>\n",
       "      <td>would realli like baggag yesterday employe akron give tough time</td>\n",
       "      <td>[would, realli, like, baggag, yesterday, employe, akron, give, tough, time]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6845</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>18</td>\n",
       "      <td>flight leav cvg gate agent no comput count passeng paper feel safe</td>\n",
       "      <td>[flight, leav, cvg, gate, agent, no, comput, count, passeng, paper, feel, safe]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10426</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>22</td>\n",
       "      <td>tri onlin told must call modif reserv hold almost hour</td>\n",
       "      <td>[tri, onlin, told, must, call, modif, reserv, hold, almost, hour]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10421</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>18</td>\n",
       "      <td>hooray hold hour never love southwest</td>\n",
       "      <td>[hooray, hold, hour, never, love, southwest]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6395</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>19</td>\n",
       "      <td>educ staff need also turn away first class cabin passeng mileag plu</td>\n",
       "      <td>[educ, staff, need, also, turn, away, first, class, cabin, passeng, mileag, plu]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10578</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>21</td>\n",
       "      <td>unfortun not think watch take bu rebook late flightr flight planestrainsandautomobil</td>\n",
       "      <td>[unfortun, not, think, watch, take, bu, rebook, late, flightr, flight, planestrainsandautomobil]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11036</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>26</td>\n",
       "      <td>peopl counter help see nobodi flight austin still sit</td>\n",
       "      <td>[peopl, counter, help, see, nobodi, flight, austin, still, sit]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13805</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>20</td>\n",
       "      <td>realli look forward flight let know reschedul diehardvirgin</td>\n",
       "      <td>[realli, look, forward, flight, let, know, reschedul, diehardvirgin]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13572</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>24</td>\n",
       "      <td>pwm direct orlando saturday happen march thought go perman thing</td>\n",
       "      <td>[pwm, direct, orlando, saturday, happen, march, thought, go, perman, thing]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5808</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "      <td>pet allow first class</td>\n",
       "      <td>[pet, allow, first, class]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>279</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>thank</td>\n",
       "      <td>[thank]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6051</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>17</td>\n",
       "      <td>would nice app would actual inform flight delay</td>\n",
       "      <td>[would, nice, app, would, actual, inform, flight, delay]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2114</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>16</td>\n",
       "      <td>give email address send actual screen shot</td>\n",
       "      <td>[give, email, address, send, actual, screen, shot]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13733</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "      <td>easi</td>\n",
       "      <td>[easi]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9454</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>8</td>\n",
       "      <td>thank info alreadi knew</td>\n",
       "      <td>[thank, info, alreadi, knew]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6235</th>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>27</td>\n",
       "      <td>fact see black histori month month year honor econom base corp like futur</td>\n",
       "      <td>[fact, see, black, histori, month, month, year, honor, econom, base, corp, like, futur]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9638</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "      <td>delay flight</td>\n",
       "      <td>[delay, flight]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>63</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "      <td>damn weather mess everyth serenitynow</td>\n",
       "      <td>[damn, weather, mess, everyth, serenitynow]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11805</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>23</td>\n",
       "      <td>ye horribl snow serious not hire peopl answer phone competitor call back number</td>\n",
       "      <td>[ye, horribl, snow, serious, not, hire, peopl, answer, phone, competitor, call, back, number]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10594</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>26</td>\n",
       "      <td>complet lack faith compani realli shame think tell challeng</td>\n",
       "      <td>[complet, lack, faith, compani, realli, shame, think, tell, challeng]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12777</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>24</td>\n",
       "      <td>spoke alreadi alreadi book anoth flight someon els thank concern want</td>\n",
       "      <td>[spoke, alreadi, alreadi, book, anoth, flight, someon, els, thank, concern, want]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11595</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>20</td>\n",
       "      <td>mail adress send ceo mr doug parker letter hope read</td>\n",
       "      <td>[mail, adress, send, ceo, mr, doug, parker, letter, hope, read]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13122</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>30</td>\n",
       "      <td>go homeless place stay tonight respond go back sleep</td>\n",
       "      <td>[go, homeless, place, stay, tonight, respond, go, back, sleep]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2375</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>19</td>\n",
       "      <td>amazingli awesom custom servic reserv agent tonight help correct mistak love airlin</td>\n",
       "      <td>[amazingli, awesom, custom, servic, reserv, agent, tonight, help, correct, mistak, love, airlin]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8257</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>20</td>\n",
       "      <td>realli asham entir busi class cabin empti almost first class</td>\n",
       "      <td>[realli, asham, entir, busi, class, cabin, empti, almost, first, class]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9559</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>18</td>\n",
       "      <td>issu mobil board pass app keep crash tri access thank</td>\n",
       "      <td>[issu, mobil, board, pass, app, keep, crash, tri, access, thank]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11689</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>28</td>\n",
       "      <td>wife group prev flight got bump reason alon two kid least put group</td>\n",
       "      <td>[wife, group, prev, flight, got, bump, reason, alon, two, kid, least, put, group]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3933</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>14</td>\n",
       "      <td>give direct number call regard lost baggag immedi</td>\n",
       "      <td>[give, direct, number, call, regard, lost, baggag, immedi]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3089</th>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>24</td>\n",
       "      <td>us return gate get mel sticker oh look heard pilot no clue mel sticker anyon</td>\n",
       "      <td>[us, return, gate, get, mel, sticker, oh, look, heard, pilot, no, clue, mel, sticker, anyon]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11579</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>19</td>\n",
       "      <td>final hung hour not abl get flight need stuckintampa</td>\n",
       "      <td>[final, hung, hour, not, abl, get, flight, need, stuckintampa]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8987</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>13</td>\n",
       "      <td>not sure</td>\n",
       "      <td>[not, sure]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4731</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>25</td>\n",
       "      <td>thank ye order rum coke get hotel voucher floor chair cushi</td>\n",
       "      <td>[thank, ye, order, rum, coke, get, hotel, voucher, floor, chair, cushi]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1464 rows × 9 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       count_capital_words  count_emojis  count_excl_quest_marks  \\\n",
       "10620  0                    0             0                        \n",
       "13609  0                    0             0                        \n",
       "2955   0                    0             0                        \n",
       "3585   0                    0             1                        \n",
       "2087   0                    0             0                        \n",
       "12382  0                    0             1                        \n",
       "4062   0                    0             0                        \n",
       "3567   0                    0             1                        \n",
       "10064  0                    0             1                        \n",
       "1809   0                    0             0                        \n",
       "14334  1                    0             0                        \n",
       "10175  2                    0             0                        \n",
       "3993   0                    0             0                        \n",
       "7851   0                    0             1                        \n",
       "1783   0                    0             1                        \n",
       "12791  2                    0             0                        \n",
       "14172  1                    0             0                        \n",
       "3878   0                    0             1                        \n",
       "12711  0                    0             1                        \n",
       "11379  0                    0             0                        \n",
       "9061   0                    0             2                        \n",
       "11739  0                    0             0                        \n",
       "13674  0                    0             0                        \n",
       "533    0                    0             0                        \n",
       "9701   0                    0             0                        \n",
       "4603   2                    0             2                        \n",
       "5835   0                    0             0                        \n",
       "4027   1                    0             0                        \n",
       "10554  0                    0             0                        \n",
       "6845   1                    0             0                        \n",
       "...   ..                   ..            ..                        \n",
       "10426  0                    0             0                        \n",
       "10421  0                    0             1                        \n",
       "6395   0                    0             0                        \n",
       "10578  0                    0             0                        \n",
       "11036  0                    0             0                        \n",
       "13805  0                    0             1                        \n",
       "13572  1                    0             1                        \n",
       "5808   0                    0             1                        \n",
       "279    0                    0             0                        \n",
       "6051   0                    0             0                        \n",
       "2114   0                    0             0                        \n",
       "13733  0                    0             0                        \n",
       "9454   0                    0             0                        \n",
       "6235   2                    0             2                        \n",
       "9638   0                    0             0                        \n",
       "63     0                    0             0                        \n",
       "11805  0                    0             1                        \n",
       "10594  0                    0             0                        \n",
       "12777  0                    0             0                        \n",
       "11595  1                    0             1                        \n",
       "13122  0                    0             2                        \n",
       "2375   0                    0             0                        \n",
       "8257   0                    0             0                        \n",
       "9559   0                    0             2                        \n",
       "11689  0                    0             0                        \n",
       "3933   0                    0             0                        \n",
       "3089   3                    0             2                        \n",
       "11579  0                    1             0                        \n",
       "8987   0                    0             0                        \n",
       "4731   0                    0             3                        \n",
       "\n",
       "       count_hashtags  count_mentions  count_urls  count_words  \\\n",
       "10620  0               1               0           6             \n",
       "13609  0               1               0           11            \n",
       "2955   0               1               0           26            \n",
       "3585   0               1               0           9             \n",
       "2087   0               1               0           3             \n",
       "12382  0               4               0           22            \n",
       "4062   0               1               0           17            \n",
       "3567   0               1               0           26            \n",
       "10064  0               1               0           11            \n",
       "1809   1               1               0           7             \n",
       "14334  1               1               0           24            \n",
       "10175  0               1               0           26            \n",
       "3993   0               1               1           6             \n",
       "7851   0               1               0           16            \n",
       "1783   0               1               0           11            \n",
       "12791  2               1               0           20            \n",
       "14172  0               1               0           25            \n",
       "3878   0               1               0           16            \n",
       "12711  0               1               0           13            \n",
       "11379  0               1               0           18            \n",
       "9061   0               1               0           16            \n",
       "11739  0               1               0           16            \n",
       "13674  0               1               0           25            \n",
       "533    0               1               0           2             \n",
       "9701   0               1               0           22            \n",
       "4603   0               4               1           19            \n",
       "5835   0               1               0           26            \n",
       "4027   0               1               0           26            \n",
       "10554  0               1               0           16            \n",
       "6845   1               1               0           18            \n",
       "...   ..              ..              ..           ..            \n",
       "10426  0               1               0           22            \n",
       "10421  0               1               0           18            \n",
       "6395   0               1               0           19            \n",
       "10578  1               1               0           21            \n",
       "11036  0               1               0           26            \n",
       "13805  1               1               0           20            \n",
       "13572  0               1               0           24            \n",
       "5808   0               1               0           7             \n",
       "279    0               1               0           2             \n",
       "6051   0               1               0           17            \n",
       "2114   0               1               0           16            \n",
       "13733  0               1               0           6             \n",
       "9454   0               1               0           8             \n",
       "6235   0               1               0           27            \n",
       "9638   0               1               0           7             \n",
       "63     1               1               0           7             \n",
       "11805  0               1               0           23            \n",
       "10594  0               2               0           26            \n",
       "12777  0               1               0           24            \n",
       "11595  0               1               0           20            \n",
       "13122  0               1               0           30            \n",
       "2375   0               1               0           19            \n",
       "8257   0               1               0           20            \n",
       "9559   0               1               0           18            \n",
       "11689  0               1               0           28            \n",
       "3933   0               1               0           14            \n",
       "3089   0               1               0           24            \n",
       "11579  1               1               0           19            \n",
       "8987   0               1               1           13            \n",
       "4731   0               1               0           25            \n",
       "\n",
       "                                                                                        clean_text  \\\n",
       "10620  chantilli pari franc                                                                          \n",
       "13609  custom servic experi tweet qualiti assur                                                      \n",
       "2955   pl stop send text everi min say flight delay anoth min hr usual defend lame                   \n",
       "3585   absolut send                                                                                  \n",
       "2087   come scoop                                                                                    \n",
       "12382  help wrong stuck florida day need learn real airlin                                           \n",
       "4062   thank find lower fare alreadi carrier direct flight juli                                      \n",
       "3567   check right hr mark board group kid age amp dnt want split sld                                \n",
       "10064  thank good back safe see soon                                                                 \n",
       "1809   link flyfi                                                                                    \n",
       "14334  unrival access california flight uk airport nonstop one stop connect year round               \n",
       "10175  thank book spent extra hr ten no explan fli middl seat char instead memphi                    \n",
       "3993   like                                                                                          \n",
       "7851   thank help total appreci made super easi                                                      \n",
       "1783   good grief flight cancel flightl hold sinc                                                    \n",
       "12791  airlin cancel flight oc fll flight today switch no notic no apolog done fli unitedairlin      \n",
       "14172  next time flight cancel flightl appreci indic email text carrier pigeon app even warn check   \n",
       "3878   tri check tomorrow flight websit keep time                                                    \n",
       "12711  access dial make faster                                                                       \n",
       "11379  right good caught mechan failur bad no blanket ok                                             \n",
       "9061   no done yet someth onlin thx                                                                  \n",
       "11739  saw one useless tsa agent inspect shoe                                                        \n",
       "13674  first good idea thatthi would happen not cancel flight earlier amateurish no plan             \n",
       "533    thank                                                                                         \n",
       "9701   morn boston beauti san diego realli bum satellit broken                                       \n",
       "4603   departur sure cater strike jfk                                                                \n",
       "5835   guess wait anoth hr hope rate goe amp reason finger cross cut close travel date               \n",
       "4027   book time call told confirm link even make chang rez                                          \n",
       "10554  would realli like baggag yesterday employe akron give tough time                              \n",
       "6845   flight leav cvg gate agent no comput count passeng paper feel safe                            \n",
       "...                                                                   ...                            \n",
       "10426  tri onlin told must call modif reserv hold almost hour                                        \n",
       "10421  hooray hold hour never love southwest                                                         \n",
       "6395   educ staff need also turn away first class cabin passeng mileag plu                           \n",
       "10578  unfortun not think watch take bu rebook late flightr flight planestrainsandautomobil          \n",
       "11036  peopl counter help see nobodi flight austin still sit                                         \n",
       "13805  realli look forward flight let know reschedul diehardvirgin                                   \n",
       "13572  pwm direct orlando saturday happen march thought go perman thing                              \n",
       "5808   pet allow first class                                                                         \n",
       "279    thank                                                                                         \n",
       "6051   would nice app would actual inform flight delay                                               \n",
       "2114   give email address send actual screen shot                                                    \n",
       "13733  easi                                                                                          \n",
       "9454   thank info alreadi knew                                                                       \n",
       "6235   fact see black histori month month year honor econom base corp like futur                     \n",
       "9638   delay flight                                                                                  \n",
       "63     damn weather mess everyth serenitynow                                                         \n",
       "11805  ye horribl snow serious not hire peopl answer phone competitor call back number               \n",
       "10594  complet lack faith compani realli shame think tell challeng                                   \n",
       "12777  spoke alreadi alreadi book anoth flight someon els thank concern want                         \n",
       "11595  mail adress send ceo mr doug parker letter hope read                                          \n",
       "13122  go homeless place stay tonight respond go back sleep                                          \n",
       "2375   amazingli awesom custom servic reserv agent tonight help correct mistak love airlin           \n",
       "8257   realli asham entir busi class cabin empti almost first class                                  \n",
       "9559   issu mobil board pass app keep crash tri access thank                                         \n",
       "11689  wife group prev flight got bump reason alon two kid least put group                           \n",
       "3933   give direct number call regard lost baggag immedi                                             \n",
       "3089   us return gate get mel sticker oh look heard pilot no clue mel sticker anyon                  \n",
       "11579  final hung hour not abl get flight need stuckintampa                                          \n",
       "8987   not sure                                                                                      \n",
       "4731   thank ye order rum coke get hotel voucher floor chair cushi                                   \n",
       "\n",
       "                                                                                               clean_text_wordlist  \n",
       "10620  [chantilli, pari, franc]                                                                                     \n",
       "13609  [custom, servic, experi, tweet, qualiti, assur]                                                              \n",
       "2955   [pl, stop, send, text, everi, min, say, flight, delay, anoth, min, hr, usual, defend, lame]                  \n",
       "3585   [absolut, send]                                                                                              \n",
       "2087   [come, scoop]                                                                                                \n",
       "12382  [help, wrong, stuck, florida, day, need, learn, real, airlin]                                                \n",
       "4062   [thank, find, lower, fare, alreadi, carrier, direct, flight, juli]                                           \n",
       "3567   [check, right, hr, mark, board, group, kid, age, amp, dnt, want, split, sld]                                 \n",
       "10064  [thank, good, back, safe, see, soon]                                                                         \n",
       "1809   [link, flyfi]                                                                                                \n",
       "14334  [unrival, access, california, flight, uk, airport, nonstop, one, stop, connect, year, round]                 \n",
       "10175  [thank, book, spent, extra, hr, ten, no, explan, fli, middl, seat, char, instead, memphi]                    \n",
       "3993   [like]                                                                                                       \n",
       "7851   [thank, help, total, appreci, made, super, easi]                                                             \n",
       "1783   [good, grief, flight, cancel, flightl, hold, sinc]                                                           \n",
       "12791  [airlin, cancel, flight, oc, fll, flight, today, switch, no, notic, no, apolog, done, fli, unitedairlin]     \n",
       "14172  [next, time, flight, cancel, flightl, appreci, indic, email, text, carrier, pigeon, app, even, warn, check]  \n",
       "3878   [tri, check, tomorrow, flight, websit, keep, time]                                                           \n",
       "12711  [access, dial, make, faster]                                                                                 \n",
       "11379  [right, good, caught, mechan, failur, bad, no, blanket, ok]                                                  \n",
       "9061   [no, done, yet, someth, onlin, thx]                                                                          \n",
       "11739  [saw, one, useless, tsa, agent, inspect, shoe]                                                               \n",
       "13674  [first, good, idea, thatthi, would, happen, not, cancel, flight, earlier, amateurish, no, plan]              \n",
       "533    [thank]                                                                                                      \n",
       "9701   [morn, boston, beauti, san, diego, realli, bum, satellit, broken]                                            \n",
       "4603   [departur, sure, cater, strike, jfk]                                                                         \n",
       "5835   [guess, wait, anoth, hr, hope, rate, goe, amp, reason, finger, cross, cut, close, travel, date]              \n",
       "4027   [book, time, call, told, confirm, link, even, make, chang, rez]                                              \n",
       "10554  [would, realli, like, baggag, yesterday, employe, akron, give, tough, time]                                  \n",
       "6845   [flight, leav, cvg, gate, agent, no, comput, count, passeng, paper, feel, safe]                              \n",
       "...                                                                                ...                              \n",
       "10426  [tri, onlin, told, must, call, modif, reserv, hold, almost, hour]                                            \n",
       "10421  [hooray, hold, hour, never, love, southwest]                                                                 \n",
       "6395   [educ, staff, need, also, turn, away, first, class, cabin, passeng, mileag, plu]                             \n",
       "10578  [unfortun, not, think, watch, take, bu, rebook, late, flightr, flight, planestrainsandautomobil]             \n",
       "11036  [peopl, counter, help, see, nobodi, flight, austin, still, sit]                                              \n",
       "13805  [realli, look, forward, flight, let, know, reschedul, diehardvirgin]                                         \n",
       "13572  [pwm, direct, orlando, saturday, happen, march, thought, go, perman, thing]                                  \n",
       "5808   [pet, allow, first, class]                                                                                   \n",
       "279    [thank]                                                                                                      \n",
       "6051   [would, nice, app, would, actual, inform, flight, delay]                                                     \n",
       "2114   [give, email, address, send, actual, screen, shot]                                                           \n",
       "13733  [easi]                                                                                                       \n",
       "9454   [thank, info, alreadi, knew]                                                                                 \n",
       "6235   [fact, see, black, histori, month, month, year, honor, econom, base, corp, like, futur]                      \n",
       "9638   [delay, flight]                                                                                              \n",
       "63     [damn, weather, mess, everyth, serenitynow]                                                                  \n",
       "11805  [ye, horribl, snow, serious, not, hire, peopl, answer, phone, competitor, call, back, number]                \n",
       "10594  [complet, lack, faith, compani, realli, shame, think, tell, challeng]                                        \n",
       "12777  [spoke, alreadi, alreadi, book, anoth, flight, someon, els, thank, concern, want]                            \n",
       "11595  [mail, adress, send, ceo, mr, doug, parker, letter, hope, read]                                              \n",
       "13122  [go, homeless, place, stay, tonight, respond, go, back, sleep]                                               \n",
       "2375   [amazingli, awesom, custom, servic, reserv, agent, tonight, help, correct, mistak, love, airlin]             \n",
       "8257   [realli, asham, entir, busi, class, cabin, empti, almost, first, class]                                      \n",
       "9559   [issu, mobil, board, pass, app, keep, crash, tri, access, thank]                                             \n",
       "11689  [wife, group, prev, flight, got, bump, reason, alon, two, kid, least, put, group]                            \n",
       "3933   [give, direct, number, call, regard, lost, baggag, immedi]                                                   \n",
       "3089   [us, return, gate, get, mel, sticker, oh, look, heard, pilot, no, clue, mel, sticker, anyon]                 \n",
       "11579  [final, hung, hour, not, abl, get, flight, need, stuckintampa]                                               \n",
       "8987   [not, sure]                                                                                                  \n",
       "4731   [thank, ye, order, rum, coke, get, hotel, voucher, floor, chair, cushi]                                      \n",
       "\n",
       "[1464 rows x 9 columns]"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test\n"
   ]
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python [default]",
   "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": "3.5.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}