EDA.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "specialized-batman",
   "metadata": {},
   "source": [
    "## 1. Imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "alternative-board",
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import nltk\n",
    "import json\n",
    "\n",
    "\n",
    "from tqdm import tqdm\n",
    "from nltk.corpus import stopwords\n",
    "from nltk.stem import *\n",
    "from nltk.stem.snowball import SnowballStemmer\n",
    "\n",
    "\n",
    "tqdm.pandas()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "talented-chassis",
   "metadata": {},
   "source": [
    "## 2. Load data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "scenic-workplace",
   "metadata": {},
   "outputs": [],
   "source": [
    "PATH_DATA = 'Data/twitter-datasets/'\n",
    "PATH = 'DATA/'\n",
    "\n",
    "t_pos = pd.read_fwf(PATH_DATA + 'train_pos_full.txt', header = None, names = ['Tweet'], colspecs = [(0,280)])\n",
    "t_neg = pd.read_fwf(PATH_DATA + 'train_neg_full.txt', header = None, names = ['Tweet'], colspecs = [(0,280)])\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "abstract-riverside",
   "metadata": {},
   "source": [
    "## 3. Counting number of words "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "lesser-interview",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 1250000/1250000 [00:03<00:00, 392967.85it/s]\n"
     ]
    }
   ],
   "source": [
    "t_pos['count_word'] = t_pos['Tweet'].progress_apply(lambda x : len(x.split()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "accompanied-period",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 1250000/1250000 [00:03<00:00, 395881.95it/s]\n"
     ]
    }
   ],
   "source": [
    "t_neg['count_word'] = t_neg['Tweet'].progress_apply(lambda x : len(x.split()))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "departmental-trademark",
   "metadata": {},
   "source": [
    "## 4. Counting number of characters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "trying-material",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 1250000/1250000 [00:01<00:00, 741397.20it/s]\n"
     ]
    }
   ],
   "source": [
    "t_pos['count_characters'] = t_pos['Tweet'].progress_apply(lambda x : len(x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "stuck-royal",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 1250000/1250000 [00:01<00:00, 722642.44it/s]\n"
     ]
    }
   ],
   "source": [
    "t_neg['count_characters'] = t_neg['Tweet'].progress_apply(lambda x : len(x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "correct-stake",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Tweet</th>\n",
       "      <th>count_word</th>\n",
       "      <th>count_characters</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>&lt;user&gt; i dunno justin read my mention or not ....</td>\n",
       "      <td>27</td>\n",
       "      <td>127</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>because your logic is so dumb , i won't even c...</td>\n",
       "      <td>21</td>\n",
       "      <td>91</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>\" &lt;user&gt; just put casper in a box ! \" looved t...</td>\n",
       "      <td>15</td>\n",
       "      <td>69</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>&lt;user&gt; &lt;user&gt; thanks sir &gt; &gt; don't trip lil ma...</td>\n",
       "      <td>17</td>\n",
       "      <td>78</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>visiting my brother tmr is the bestest birthda...</td>\n",
       "      <td>13</td>\n",
       "      <td>66</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                               Tweet  count_word  \\\n",
       "0  <user> i dunno justin read my mention or not ....          27   \n",
       "1  because your logic is so dumb , i won't even c...          21   \n",
       "2  \" <user> just put casper in a box ! \" looved t...          15   \n",
       "3  <user> <user> thanks sir > > don't trip lil ma...          17   \n",
       "4  visiting my brother tmr is the bestest birthda...          13   \n",
       "\n",
       "   count_characters  \n",
       "0               127  \n",
       "1                91  \n",
       "2                69  \n",
       "3                78  \n",
       "4                66  "
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "t_pos.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "adverse-killer",
   "metadata": {},
   "source": [
    "## 5. Replacing contractions and lowercasing and stemming\n",
    "https://stackoverflow.com/questions/46786211/counting-the-frequency-of-words-in-a-pandas-data-frame"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "divided-spokesman",
   "metadata": {},
   "outputs": [],
   "source": [
    "t_pos['Tweet'] = t_pos['Tweet'].apply(lambda x : x.lower())\n",
    "t_neg['Tweet'] = t_neg['Tweet'].apply(lambda x : x.lower())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "recognized-purse",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Tweet</th>\n",
       "      <th>count_word</th>\n",
       "      <th>count_characters</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>&lt;user&gt; i dunno justin read my mention or not ....</td>\n",
       "      <td>27</td>\n",
       "      <td>127</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>because your logic is so dumb , i won't even c...</td>\n",
       "      <td>21</td>\n",
       "      <td>91</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>\" &lt;user&gt; just put casper in a box ! \" looved t...</td>\n",
       "      <td>15</td>\n",
       "      <td>69</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>&lt;user&gt; &lt;user&gt; thanks sir &gt; &gt; don't trip lil ma...</td>\n",
       "      <td>17</td>\n",
       "      <td>78</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>visiting my brother tmr is the bestest birthda...</td>\n",
       "      <td>13</td>\n",
       "      <td>66</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>95</th>\n",
       "      <td>&lt;user&gt; it's once every 24 hours</td>\n",
       "      <td>6</td>\n",
       "      <td>31</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>96</th>\n",
       "      <td>&lt;user&gt; follow &lt;user&gt; cos she's a fully sick dog</td>\n",
       "      <td>9</td>\n",
       "      <td>47</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>97</th>\n",
       "      <td>&lt;user&gt; &lt;user&gt; thanks for the super sexy gold s...</td>\n",
       "      <td>15</td>\n",
       "      <td>78</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>98</th>\n",
       "      <td>&lt;user&gt; &lt;user&gt; &lt;user&gt; &lt;user&gt; &lt;user&gt; &lt;user&gt; &lt;use...</td>\n",
       "      <td>11</td>\n",
       "      <td>72</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99</th>\n",
       "      <td>&lt;user&gt; i'm not sure if they are really aware o...</td>\n",
       "      <td>21</td>\n",
       "      <td>103</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>100 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                Tweet  count_word  \\\n",
       "0   <user> i dunno justin read my mention or not ....          27   \n",
       "1   because your logic is so dumb , i won't even c...          21   \n",
       "2   \" <user> just put casper in a box ! \" looved t...          15   \n",
       "3   <user> <user> thanks sir > > don't trip lil ma...          17   \n",
       "4   visiting my brother tmr is the bestest birthda...          13   \n",
       "..                                                ...         ...   \n",
       "95                    <user> it's once every 24 hours           6   \n",
       "96    <user> follow <user> cos she's a fully sick dog           9   \n",
       "97  <user> <user> thanks for the super sexy gold s...          15   \n",
       "98  <user> <user> <user> <user> <user> <user> <use...          11   \n",
       "99  <user> i'm not sure if they are really aware o...          21   \n",
       "\n",
       "    count_characters  \n",
       "0                127  \n",
       "1                 91  \n",
       "2                 69  \n",
       "3                 78  \n",
       "4                 66  \n",
       "..               ...  \n",
       "95                31  \n",
       "96                47  \n",
       "97                78  \n",
       "98                72  \n",
       "99               103  \n",
       "\n",
       "[100 rows x 3 columns]"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "t_pos.head(100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "forty-narrative",
   "metadata": {},
   "outputs": [],
   "source": [
    "def load_contractions():\n",
    "    contraction_list = json.loads(open(PATH + 'english_contractions.json', 'r').read())\n",
    "    contraction_list['ur'] = \"your\"\n",
    "    contraction_list['ya'] = \"you\"\n",
    "    return contraction_list\n",
    "\n",
    "def remove_contractions(sentence, contraction_list):\n",
    "    sent_split = sentence.split(\" \")\n",
    "    sent_wo = [contraction_list[word] if word in contraction_list else word for word in sent_split]\n",
    "    return \" \".join(sent_wo)\n",
    "\n",
    "def remove_contractions_df(df, contraction_list,  col = 'Tweet'):\n",
    "    df[col + '_cont'] = df[col].progress_apply(lambda x : remove_contractions(x, contraction_list))\n",
    "    return df "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "destroyed-insight",
   "metadata": {},
   "outputs": [],
   "source": [
    "slang_list = json.loads(open(PATH + 'slang.json', 'r').read())\n",
    "\n",
    "def remove_slang(sentence):\n",
    "    sent_split = sentence.split(\" \")\n",
    "    sent_wo = [slang_list[word] if word in slang_list else word for word in sent_split]\n",
    "    return \" \".join(sent_wo)\n",
    "\n",
    "def remove_slang_df(df, col = 'Tweet_cont'):\n",
    "    df[col] = df[col].progress_apply(lambda x : remove_slang(x))\n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "blocked-syracuse",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 1250000/1250000 [00:06<00:00, 197524.34it/s]\n",
      "100%|██████████| 1250000/1250000 [00:06<00:00, 182064.62it/s]\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Tweet</th>\n",
       "      <th>count_word</th>\n",
       "      <th>count_characters</th>\n",
       "      <th>Tweet_cont</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>vinco tresorpack 6 ( difficulty 10 of 10 objec...</td>\n",
       "      <td>21</td>\n",
       "      <td>122</td>\n",
       "      <td>vinco tresorpack 6 ( difficulty 10 of 10 objec...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>glad i dot have taks tomorrow ! ! #thankful #s...</td>\n",
       "      <td>10</td>\n",
       "      <td>52</td>\n",
       "      <td>glad i dot have taks tomorrow ! ! #thankful #s...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1-3 vs celtics in the regular season = were fu...</td>\n",
       "      <td>17</td>\n",
       "      <td>82</td>\n",
       "      <td>1-3 vs celtics in the regular season = we are ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>&lt;user&gt; i could actually kill that girl i'm so ...</td>\n",
       "      <td>13</td>\n",
       "      <td>57</td>\n",
       "      <td>&lt;user&gt; i could actually kill that girl I am so...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>&lt;user&gt; &lt;user&gt; &lt;user&gt; i find that very hard to ...</td>\n",
       "      <td>12</td>\n",
       "      <td>63</td>\n",
       "      <td>&lt;user&gt; &lt;user&gt; &lt;user&gt; i find that very hard to ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1249995</th>\n",
       "      <td>im so sorry ! &lt;user&gt; &amp; to &lt;user&gt; &amp; &lt;user&gt; u gu...</td>\n",
       "      <td>16</td>\n",
       "      <td>69</td>\n",
       "      <td>I am so sorry ! &lt;user&gt; &amp; to &lt;user&gt; &amp; &lt;user&gt; u ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1249996</th>\n",
       "      <td>i can't find food coloring anywhere</td>\n",
       "      <td>6</td>\n",
       "      <td>35</td>\n",
       "      <td>i cannot find food coloring anywhere</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1249997</th>\n",
       "      <td>&lt;user&gt; same here ! ! but tort ! ! wonder why y...</td>\n",
       "      <td>27</td>\n",
       "      <td>131</td>\n",
       "      <td>&lt;user&gt; same here ! ! but tort ! ! wonder why y...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1249998</th>\n",
       "      <td>keyless entry remote fob clicker for 2005 buic...</td>\n",
       "      <td>22</td>\n",
       "      <td>120</td>\n",
       "      <td>keyless entry remote fob clicker for 2005 buic...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1249999</th>\n",
       "      <td>&lt;user&gt; yeap . doctor don't know what's wrong w...</td>\n",
       "      <td>26</td>\n",
       "      <td>127</td>\n",
       "      <td>&lt;user&gt; yeap . doctor do not know what has wron...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1250000 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                     Tweet  count_word  \\\n",
       "0        vinco tresorpack 6 ( difficulty 10 of 10 objec...          21   \n",
       "1        glad i dot have taks tomorrow ! ! #thankful #s...          10   \n",
       "2        1-3 vs celtics in the regular season = were fu...          17   \n",
       "3        <user> i could actually kill that girl i'm so ...          13   \n",
       "4        <user> <user> <user> i find that very hard to ...          12   \n",
       "...                                                    ...         ...   \n",
       "1249995  im so sorry ! <user> & to <user> & <user> u gu...          16   \n",
       "1249996                i can't find food coloring anywhere           6   \n",
       "1249997  <user> same here ! ! but tort ! ! wonder why y...          27   \n",
       "1249998  keyless entry remote fob clicker for 2005 buic...          22   \n",
       "1249999  <user> yeap . doctor don't know what's wrong w...          26   \n",
       "\n",
       "         count_characters                                         Tweet_cont  \n",
       "0                     122  vinco tresorpack 6 ( difficulty 10 of 10 objec...  \n",
       "1                      52  glad i dot have taks tomorrow ! ! #thankful #s...  \n",
       "2                      82  1-3 vs celtics in the regular season = we are ...  \n",
       "3                      57  <user> i could actually kill that girl I am so...  \n",
       "4                      63  <user> <user> <user> i find that very hard to ...  \n",
       "...                   ...                                                ...  \n",
       "1249995                69  I am so sorry ! <user> & to <user> & <user> u ...  \n",
       "1249996                35               i cannot find food coloring anywhere  \n",
       "1249997               131  <user> same here ! ! but tort ! ! wonder why y...  \n",
       "1249998               120  keyless entry remote fob clicker for 2005 buic...  \n",
       "1249999               127  <user> yeap . doctor do not know what has wron...  \n",
       "\n",
       "[1250000 rows x 4 columns]"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "contract_ls = load_contractions()\n",
    "\n",
    "remove_contractions_df(t_pos, contract_ls)\n",
    "remove_contractions_df(t_neg, contract_ls)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "complex-charlotte",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0           True\n",
       "1           True\n",
       "2           True\n",
       "3           True\n",
       "4           True\n",
       "           ...  \n",
       "1249995     True\n",
       "1249996     True\n",
       "1249997     True\n",
       "1249998     True\n",
       "1249999    False\n",
       "Name: Tweet_cont, Length: 1250000, dtype: bool"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "t_pos.Tweet_cont.str.islower()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "ambient-moment",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'<user> I am so excited for tomorrow ! look out for two leprechauns ! xx'"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "t_pos.Tweet_cont.iloc[99998]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "attended-hanging",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 1250000/1250000 [00:06<00:00, 181486.88it/s]\n",
      "100%|██████████| 1250000/1250000 [00:08<00:00, 148618.21it/s]\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Tweet</th>\n",
       "      <th>count_word</th>\n",
       "      <th>count_characters</th>\n",
       "      <th>Tweet_cont</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>vinco tresorpack 6 ( difficulty 10 of 10 objec...</td>\n",
       "      <td>21</td>\n",
       "      <td>122</td>\n",
       "      <td>vinco tresorpack 6 ( difficulty 10 of 10 objec...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>glad i dot have taks tomorrow ! ! #thankful #s...</td>\n",
       "      <td>10</td>\n",
       "      <td>52</td>\n",
       "      <td>glad i dot have taks tomorrow ! ! #thankful #s...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1-3 vs celtics in the regular season = were fu...</td>\n",
       "      <td>17</td>\n",
       "      <td>82</td>\n",
       "      <td>1-3 vs celtics in the regular season = we are ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>&lt;user&gt; i could actually kill that girl i'm so ...</td>\n",
       "      <td>13</td>\n",
       "      <td>57</td>\n",
       "      <td>&lt;user&gt; i could actually kill that girl I am so...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>&lt;user&gt; &lt;user&gt; &lt;user&gt; i find that very hard to ...</td>\n",
       "      <td>12</td>\n",
       "      <td>63</td>\n",
       "      <td>&lt;user&gt; &lt;user&gt; &lt;user&gt; i find that very hard to ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1249995</th>\n",
       "      <td>im so sorry ! &lt;user&gt; &amp; to &lt;user&gt; &amp; &lt;user&gt; u gu...</td>\n",
       "      <td>16</td>\n",
       "      <td>69</td>\n",
       "      <td>I am so sorry ! &lt;user&gt; &amp; to &lt;user&gt; &amp; &lt;user&gt; yo...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1249996</th>\n",
       "      <td>i can't find food coloring anywhere</td>\n",
       "      <td>6</td>\n",
       "      <td>35</td>\n",
       "      <td>i cannot find food coloring anywhere</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1249997</th>\n",
       "      <td>&lt;user&gt; same here ! ! but tort ! ! wonder why y...</td>\n",
       "      <td>27</td>\n",
       "      <td>131</td>\n",
       "      <td>&lt;user&gt; same here ! ! but tort ! ! wonder why y...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1249998</th>\n",
       "      <td>keyless entry remote fob clicker for 2005 buic...</td>\n",
       "      <td>22</td>\n",
       "      <td>120</td>\n",
       "      <td>keyless entry remote fresh off the boat clicke...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1249999</th>\n",
       "      <td>&lt;user&gt; yeap . doctor don't know what's wrong w...</td>\n",
       "      <td>26</td>\n",
       "      <td>127</td>\n",
       "      <td>&lt;user&gt; yeap . doctor do not know what has wron...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1250000 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                     Tweet  count_word  \\\n",
       "0        vinco tresorpack 6 ( difficulty 10 of 10 objec...          21   \n",
       "1        glad i dot have taks tomorrow ! ! #thankful #s...          10   \n",
       "2        1-3 vs celtics in the regular season = were fu...          17   \n",
       "3        <user> i could actually kill that girl i'm so ...          13   \n",
       "4        <user> <user> <user> i find that very hard to ...          12   \n",
       "...                                                    ...         ...   \n",
       "1249995  im so sorry ! <user> & to <user> & <user> u gu...          16   \n",
       "1249996                i can't find food coloring anywhere           6   \n",
       "1249997  <user> same here ! ! but tort ! ! wonder why y...          27   \n",
       "1249998  keyless entry remote fob clicker for 2005 buic...          22   \n",
       "1249999  <user> yeap . doctor don't know what's wrong w...          26   \n",
       "\n",
       "         count_characters                                         Tweet_cont  \n",
       "0                     122  vinco tresorpack 6 ( difficulty 10 of 10 objec...  \n",
       "1                      52  glad i dot have taks tomorrow ! ! #thankful #s...  \n",
       "2                      82  1-3 vs celtics in the regular season = we are ...  \n",
       "3                      57  <user> i could actually kill that girl I am so...  \n",
       "4                      63  <user> <user> <user> i find that very hard to ...  \n",
       "...                   ...                                                ...  \n",
       "1249995                69  I am so sorry ! <user> & to <user> & <user> yo...  \n",
       "1249996                35               i cannot find food coloring anywhere  \n",
       "1249997               131  <user> same here ! ! but tort ! ! wonder why y...  \n",
       "1249998               120  keyless entry remote fresh off the boat clicke...  \n",
       "1249999               127  <user> yeap . doctor do not know what has wron...  \n",
       "\n",
       "[1250000 rows x 4 columns]"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "remove_slang_df(t_pos)\n",
    "remove_slang_df(t_neg)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "surgical-candidate",
   "metadata": {},
   "outputs": [],
   "source": [
    "t_pos['Tweet_cont'] = t_pos['Tweet_cont'].apply(lambda x : x.lower())\n",
    "t_neg['Tweet_cont'] = t_neg['Tweet_cont'].apply(lambda x : x.lower())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "medium-detective",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'<user> i am so excited for tomorrow ! look out for two leprechauns ! xx'"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "t_pos.Tweet_cont.iloc[99998]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "coated-friendly",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 1250000/1250000 [04:16<00:00, 4871.03it/s]\n",
      "100%|██████████| 1250000/1250000 [04:51<00:00, 4292.01it/s]\n"
     ]
    }
   ],
   "source": [
    "stemmer = SnowballStemmer(\"english\")\n",
    "\n",
    "def stem_word(sent):\n",
    "    sent_sp = sent.split()\n",
    "    st = [stemmer.stem(w) for w in sent_sp]\n",
    "    return \" \".join(st)\n",
    "\n",
    "t_pos['Tweet_cont'] = t_pos['Tweet_cont'].progress_apply(lambda x : stem_word(x))\n",
    "t_neg['Tweet_cont'] = t_neg['Tweet_cont'].progress_apply(lambda x : stem_word(x))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "through-logan",
   "metadata": {},
   "source": [
    "## 6. Most frequent stop-words"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "otherwise-station",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package stopwords to\n",
      "[nltk_data]     /Users/clemencebarsi/nltk_data...\n",
      "[nltk_data]   Package stopwords is already up-to-date!\n",
      "100%|██████████| 1250000/1250000 [00:13<00:00, 90505.65it/s]\n",
      "100%|██████████| 1250000/1250000 [00:13<00:00, 95163.98it/s]\n",
      "100%|██████████| 1250000/1250000 [00:14<00:00, 88171.01it/s]\n",
      "100%|██████████| 1250000/1250000 [00:14<00:00, 86328.79it/s]\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Tweet</th>\n",
       "      <th>count_word</th>\n",
       "      <th>count_characters</th>\n",
       "      <th>Tweet_cont</th>\n",
       "      <th>no_sw</th>\n",
       "      <th>sw</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>vinco tresorpack 6 ( difficulty 10 of 10 objec...</td>\n",
       "      <td>21</td>\n",
       "      <td>122</td>\n",
       "      <td>vinco tresorpack 6 ( difficulti 10 of 10 objec...</td>\n",
       "      <td>vinco tresorpack 6 ( difficulti 10 10 object :...</td>\n",
       "      <td>of and the this &lt;url&gt;</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>glad i dot have taks tomorrow ! ! #thankful #s...</td>\n",
       "      <td>10</td>\n",
       "      <td>52</td>\n",
       "      <td>glad i dot have tak tomorrow ! ! #thank #startho</td>\n",
       "      <td>glad dot tak tomorrow ! ! #thank #startho</td>\n",
       "      <td>i have</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1-3 vs celtics in the regular season = were fu...</td>\n",
       "      <td>17</td>\n",
       "      <td>82</td>\n",
       "      <td>1-3 vs celtic in the regular season = we are f...</td>\n",
       "      <td>1-3 vs celtic regular season = fuck play playoff</td>\n",
       "      <td>in the we are if we them in the</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>&lt;user&gt; i could actually kill that girl i'm so ...</td>\n",
       "      <td>13</td>\n",
       "      <td>57</td>\n",
       "      <td>&lt;user&gt; i could actual kill that girl i am so s...</td>\n",
       "      <td>could actual kill girl sorri ! ! !</td>\n",
       "      <td>&lt;user&gt; i that i am so</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>&lt;user&gt; &lt;user&gt; &lt;user&gt; i find that very hard to ...</td>\n",
       "      <td>12</td>\n",
       "      <td>63</td>\n",
       "      <td>&lt;user&gt; &lt;user&gt; &lt;user&gt; i find that veri hard to ...</td>\n",
       "      <td>find veri hard believ afraid</td>\n",
       "      <td>&lt;user&gt; &lt;user&gt; &lt;user&gt; i that to i am</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1249995</th>\n",
       "      <td>im so sorry ! &lt;user&gt; &amp; to &lt;user&gt; &amp; &lt;user&gt; u gu...</td>\n",
       "      <td>16</td>\n",
       "      <td>69</td>\n",
       "      <td>i am so sorri ! &lt;user&gt; &amp; to &lt;user&gt; &amp; &lt;user&gt; yo...</td>\n",
       "      <td>sorri ! &amp; &amp; guy ridicul ! &gt;:/</td>\n",
       "      <td>i am so &lt;user&gt; to &lt;user&gt; &lt;user&gt; you are</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1249996</th>\n",
       "      <td>i can't find food coloring anywhere</td>\n",
       "      <td>6</td>\n",
       "      <td>35</td>\n",
       "      <td>i cannot find food color anywher</td>\n",
       "      <td>cannot find food color anywher</td>\n",
       "      <td>i</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1249997</th>\n",
       "      <td>&lt;user&gt; same here ! ! but tort ! ! wonder why y...</td>\n",
       "      <td>27</td>\n",
       "      <td>131</td>\n",
       "      <td>&lt;user&gt; same here ! ! but tort ! ! wonder whi y...</td>\n",
       "      <td>! ! tort ! ! wonder whi chose crime modul haha...</td>\n",
       "      <td>&lt;user&gt; same here but you over all the other to...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1249998</th>\n",
       "      <td>keyless entry remote fob clicker for 2005 buic...</td>\n",
       "      <td>22</td>\n",
       "      <td>120</td>\n",
       "      <td>keyless entri remot fresh off the boat clicker...</td>\n",
       "      <td>keyless entri remot fresh boat clicker 2005 bu...</td>\n",
       "      <td>off the for be by not &lt;url&gt;</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1249999</th>\n",
       "      <td>&lt;user&gt; yeap . doctor don't know what's wrong w...</td>\n",
       "      <td>26</td>\n",
       "      <td>127</td>\n",
       "      <td>&lt;user&gt; yeap . doctor do not know what has wron...</td>\n",
       "      <td>yeap . doctor know wrong laugh loud , either m...</td>\n",
       "      <td>&lt;user&gt; do not what has with me out or i don't ...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1250000 rows × 6 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                     Tweet  count_word  \\\n",
       "0        vinco tresorpack 6 ( difficulty 10 of 10 objec...          21   \n",
       "1        glad i dot have taks tomorrow ! ! #thankful #s...          10   \n",
       "2        1-3 vs celtics in the regular season = were fu...          17   \n",
       "3        <user> i could actually kill that girl i'm so ...          13   \n",
       "4        <user> <user> <user> i find that very hard to ...          12   \n",
       "...                                                    ...         ...   \n",
       "1249995  im so sorry ! <user> & to <user> & <user> u gu...          16   \n",
       "1249996                i can't find food coloring anywhere           6   \n",
       "1249997  <user> same here ! ! but tort ! ! wonder why y...          27   \n",
       "1249998  keyless entry remote fob clicker for 2005 buic...          22   \n",
       "1249999  <user> yeap . doctor don't know what's wrong w...          26   \n",
       "\n",
       "         count_characters                                         Tweet_cont  \\\n",
       "0                     122  vinco tresorpack 6 ( difficulti 10 of 10 objec...   \n",
       "1                      52   glad i dot have tak tomorrow ! ! #thank #startho   \n",
       "2                      82  1-3 vs celtic in the regular season = we are f...   \n",
       "3                      57  <user> i could actual kill that girl i am so s...   \n",
       "4                      63  <user> <user> <user> i find that veri hard to ...   \n",
       "...                   ...                                                ...   \n",
       "1249995                69  i am so sorri ! <user> & to <user> & <user> yo...   \n",
       "1249996                35                   i cannot find food color anywher   \n",
       "1249997               131  <user> same here ! ! but tort ! ! wonder whi y...   \n",
       "1249998               120  keyless entri remot fresh off the boat clicker...   \n",
       "1249999               127  <user> yeap . doctor do not know what has wron...   \n",
       "\n",
       "                                                     no_sw  \\\n",
       "0        vinco tresorpack 6 ( difficulti 10 10 object :...   \n",
       "1                glad dot tak tomorrow ! ! #thank #startho   \n",
       "2         1-3 vs celtic regular season = fuck play playoff   \n",
       "3                       could actual kill girl sorri ! ! !   \n",
       "4                             find veri hard believ afraid   \n",
       "...                                                    ...   \n",
       "1249995                      sorri ! & & guy ridicul ! >:/   \n",
       "1249996                     cannot find food color anywher   \n",
       "1249997  ! ! tort ! ! wonder whi chose crime modul haha...   \n",
       "1249998  keyless entri remot fresh boat clicker 2005 bu...   \n",
       "1249999  yeap . doctor know wrong laugh loud , either m...   \n",
       "\n",
       "                                                        sw  \n",
       "0                                    of and the this <url>  \n",
       "1                                                   i have  \n",
       "2                          in the we are if we them in the  \n",
       "3                                    <user> i that i am so  \n",
       "4                      <user> <user> <user> i that to i am  \n",
       "...                                                    ...  \n",
       "1249995            i am so <user> to <user> <user> you are  \n",
       "1249996                                                  i  \n",
       "1249997  <user> same here but you over all the other to...  \n",
       "1249998                        off the for be by not <url>  \n",
       "1249999  <user> do not what has with me out or i don't ...  \n",
       "\n",
       "[1250000 rows x 6 columns]"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nltk.download('stopwords')\n",
    "\n",
    "def load_stopwords():\n",
    "    all_stopwords = stopwords.words('english')\n",
    "    all_stopwords.append('<user>')\n",
    "    all_stopwords.append('<url>')\n",
    "    all_stopwords.append('i')\n",
    "    return all_stopwords\n",
    "\n",
    "def remove_stopwords(sentence, all_stopwords):\n",
    "    sent_split = sentence.split(\" \")\n",
    "    \n",
    "    all_stopwords = set(all_stopwords) # set() heremakes it at least 10x faster\n",
    "    sent_without_sw = [word for word in sent_split if not word in all_stopwords]\n",
    "    no_sw = \" \".join(sent_without_sw)\n",
    "    return no_sw\n",
    "\n",
    "def only_stopwords(sentence, all_stopwords):\n",
    "    sent_split = sentence.split(\" \")\n",
    "    all_stopwords = set(all_stopwords)\n",
    "    sw = [word for word in sent_split if word in all_stopwords]\n",
    "    sw_j = \" \".join(sw)\n",
    "    return sw_j\n",
    "\n",
    "def remove_stopwords_df(df, col = 'Tweet_cont'):\n",
    "    df['no_sw'] = df[col].progress_apply(lambda x : remove_stopwords(x, all_stopwords))\n",
    "    df['sw'] = df[col].progress_apply(lambda x : only_stopwords(x, all_stopwords))\n",
    "    return df\n",
    "\n",
    "all_stopwords = load_stopwords()\n",
    "remove_stopwords_df(t_pos)\n",
    "remove_stopwords_df(t_neg)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "global-growth",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0          True\n",
       "1          True\n",
       "2          True\n",
       "3          True\n",
       "4          True\n",
       "           ... \n",
       "1249995    True\n",
       "1249996    True\n",
       "1249997    True\n",
       "1249998    True\n",
       "1249999    True\n",
       "Name: Tweet_cont, Length: 1250000, dtype: bool"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "t_pos.Tweet_cont.str.islower()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "naked-compatibility",
   "metadata": {},
   "outputs": [],
   "source": [
    "count_pos_sw = t_pos.sw.str.split(expand = True).stack().value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "appreciated-realtor",
   "metadata": {},
   "outputs": [],
   "source": [
    "count_neg_sw = t_neg.sw.str.split(expand = True).stack().value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "increased-locator",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<user>    1027186\n",
       "i          639229\n",
       "you        528273\n",
       "to         393353\n",
       "the        339442\n",
       "           ...   \n",
       "weren           2\n",
       "hasn            2\n",
       "hadn            1\n",
       "being           1\n",
       "you'll          1\n",
       "Length: 143, dtype: int64"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "count_pos_sw"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "complex-lightweight",
   "metadata": {},
   "source": [
    "## 7. Most frequent special characters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "alien-colleague",
   "metadata": {},
   "outputs": [],
   "source": [
    "tokenizer = nltk.RegexpTokenizer(r\"\\w+\")\n",
    "import re\n",
    "\n",
    "pattern = r'[0-9]'\n",
    "\n",
    "from string import ascii_letters, digits, punctuation\n",
    "\n",
    "\n",
    "def remove_punct_df(df, col = 'no_sw'):\n",
    "    df['without_punc'] = df[col].progress_apply(lambda x :\" \".join(tokenizer.tokenize(x)))\n",
    "    df['punc'] = df[col].progress_apply(lambda x : re.sub(r'[^{}]+'.format(punctuation),' ',x))\n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "quantitative-athens",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 1250000/1250000 [00:06<00:00, 193368.72it/s]\n",
      "100%|██████████| 1250000/1250000 [00:05<00:00, 232251.17it/s]\n",
      "100%|██████████| 1250000/1250000 [00:07<00:00, 163242.35it/s]\n",
      "100%|██████████| 1250000/1250000 [00:05<00:00, 218886.51it/s]\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Tweet</th>\n",
       "      <th>count_word</th>\n",
       "      <th>count_characters</th>\n",
       "      <th>Tweet_cont</th>\n",
       "      <th>no_sw</th>\n",
       "      <th>sw</th>\n",
       "      <th>without_punc</th>\n",
       "      <th>punc</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>vinco tresorpack 6 ( difficulty 10 of 10 objec...</td>\n",
       "      <td>21</td>\n",
       "      <td>122</td>\n",
       "      <td>vinco tresorpack 6 ( difficulti 10 of 10 objec...</td>\n",
       "      <td>vinco tresorpack 6 ( difficulti 10 10 object :...</td>\n",
       "      <td>of and the this &lt;url&gt;</td>\n",
       "      <td>vinco tresorpack 6 difficulti 10 10 object dis...</td>\n",
       "      <td>( : ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>glad i dot have taks tomorrow ! ! #thankful #s...</td>\n",
       "      <td>10</td>\n",
       "      <td>52</td>\n",
       "      <td>glad i dot have tak tomorrow ! ! #thank #startho</td>\n",
       "      <td>glad dot tak tomorrow ! ! #thank #startho</td>\n",
       "      <td>i have</td>\n",
       "      <td>glad dot tak tomorrow thank startho</td>\n",
       "      <td>! ! # #</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1-3 vs celtics in the regular season = were fu...</td>\n",
       "      <td>17</td>\n",
       "      <td>82</td>\n",
       "      <td>1-3 vs celtic in the regular season = we are f...</td>\n",
       "      <td>1-3 vs celtic regular season = fuck play playoff</td>\n",
       "      <td>in the we are if we them in the</td>\n",
       "      <td>1 3 vs celtic regular season fuck play playoff</td>\n",
       "      <td>- =</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>&lt;user&gt; i could actually kill that girl i'm so ...</td>\n",
       "      <td>13</td>\n",
       "      <td>57</td>\n",
       "      <td>&lt;user&gt; i could actual kill that girl i am so s...</td>\n",
       "      <td>could actual kill girl sorri ! ! !</td>\n",
       "      <td>&lt;user&gt; i that i am so</td>\n",
       "      <td>could actual kill girl sorri</td>\n",
       "      <td>! ! !</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>&lt;user&gt; &lt;user&gt; &lt;user&gt; i find that very hard to ...</td>\n",
       "      <td>12</td>\n",
       "      <td>63</td>\n",
       "      <td>&lt;user&gt; &lt;user&gt; &lt;user&gt; i find that veri hard to ...</td>\n",
       "      <td>find veri hard believ afraid</td>\n",
       "      <td>&lt;user&gt; &lt;user&gt; &lt;user&gt; i that to i am</td>\n",
       "      <td>find veri hard believ afraid</td>\n",
       "      <td></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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1249995</th>\n",
       "      <td>im so sorry ! &lt;user&gt; &amp; to &lt;user&gt; &amp; &lt;user&gt; u gu...</td>\n",
       "      <td>16</td>\n",
       "      <td>69</td>\n",
       "      <td>i am so sorri ! &lt;user&gt; &amp; to &lt;user&gt; &amp; &lt;user&gt; yo...</td>\n",
       "      <td>sorri ! &amp; &amp; guy ridicul ! &gt;:/</td>\n",
       "      <td>i am so &lt;user&gt; to &lt;user&gt; &lt;user&gt; you are</td>\n",
       "      <td>sorri guy ridicul</td>\n",
       "      <td>! &amp; &amp; ! &gt;:/</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1249996</th>\n",
       "      <td>i can't find food coloring anywhere</td>\n",
       "      <td>6</td>\n",
       "      <td>35</td>\n",
       "      <td>i cannot find food color anywher</td>\n",
       "      <td>cannot find food color anywher</td>\n",
       "      <td>i</td>\n",
       "      <td>cannot find food color anywher</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1249997</th>\n",
       "      <td>&lt;user&gt; same here ! ! but tort ! ! wonder why y...</td>\n",
       "      <td>27</td>\n",
       "      <td>131</td>\n",
       "      <td>&lt;user&gt; same here ! ! but tort ! ! wonder whi y...</td>\n",
       "      <td>! ! tort ! ! wonder whi chose crime modul haha...</td>\n",
       "      <td>&lt;user&gt; same here but you over all the other to...</td>\n",
       "      <td>tort wonder whi chose crime modul hahaa tri im...</td>\n",
       "      <td>! ! ! ! ?</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1249998</th>\n",
       "      <td>keyless entry remote fob clicker for 2005 buic...</td>\n",
       "      <td>22</td>\n",
       "      <td>120</td>\n",
       "      <td>keyless entri remot fresh off the boat clicker...</td>\n",
       "      <td>keyless entri remot fresh boat clicker 2005 bu...</td>\n",
       "      <td>off the for be by not &lt;url&gt;</td>\n",
       "      <td>keyless entri remot fresh boat clicker 2005 bu...</td>\n",
       "      <td>- ( ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1249999</th>\n",
       "      <td>&lt;user&gt; yeap . doctor don't know what's wrong w...</td>\n",
       "      <td>26</td>\n",
       "      <td>127</td>\n",
       "      <td>&lt;user&gt; yeap . doctor do not know what has wron...</td>\n",
       "      <td>yeap . doctor know wrong laugh loud , either m...</td>\n",
       "      <td>&lt;user&gt; do not what has with me out or i don't ...</td>\n",
       "      <td>yeap doctor know wrong laugh loud either migra...</td>\n",
       "      <td>. , . . .</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1250000 rows × 8 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                     Tweet  count_word  \\\n",
       "0        vinco tresorpack 6 ( difficulty 10 of 10 objec...          21   \n",
       "1        glad i dot have taks tomorrow ! ! #thankful #s...          10   \n",
       "2        1-3 vs celtics in the regular season = were fu...          17   \n",
       "3        <user> i could actually kill that girl i'm so ...          13   \n",
       "4        <user> <user> <user> i find that very hard to ...          12   \n",
       "...                                                    ...         ...   \n",
       "1249995  im so sorry ! <user> & to <user> & <user> u gu...          16   \n",
       "1249996                i can't find food coloring anywhere           6   \n",
       "1249997  <user> same here ! ! but tort ! ! wonder why y...          27   \n",
       "1249998  keyless entry remote fob clicker for 2005 buic...          22   \n",
       "1249999  <user> yeap . doctor don't know what's wrong w...          26   \n",
       "\n",
       "         count_characters                                         Tweet_cont  \\\n",
       "0                     122  vinco tresorpack 6 ( difficulti 10 of 10 objec...   \n",
       "1                      52   glad i dot have tak tomorrow ! ! #thank #startho   \n",
       "2                      82  1-3 vs celtic in the regular season = we are f...   \n",
       "3                      57  <user> i could actual kill that girl i am so s...   \n",
       "4                      63  <user> <user> <user> i find that veri hard to ...   \n",
       "...                   ...                                                ...   \n",
       "1249995                69  i am so sorri ! <user> & to <user> & <user> yo...   \n",
       "1249996                35                   i cannot find food color anywher   \n",
       "1249997               131  <user> same here ! ! but tort ! ! wonder whi y...   \n",
       "1249998               120  keyless entri remot fresh off the boat clicker...   \n",
       "1249999               127  <user> yeap . doctor do not know what has wron...   \n",
       "\n",
       "                                                     no_sw  \\\n",
       "0        vinco tresorpack 6 ( difficulti 10 10 object :...   \n",
       "1                glad dot tak tomorrow ! ! #thank #startho   \n",
       "2         1-3 vs celtic regular season = fuck play playoff   \n",
       "3                       could actual kill girl sorri ! ! !   \n",
       "4                             find veri hard believ afraid   \n",
       "...                                                    ...   \n",
       "1249995                      sorri ! & & guy ridicul ! >:/   \n",
       "1249996                     cannot find food color anywher   \n",
       "1249997  ! ! tort ! ! wonder whi chose crime modul haha...   \n",
       "1249998  keyless entri remot fresh boat clicker 2005 bu...   \n",
       "1249999  yeap . doctor know wrong laugh loud , either m...   \n",
       "\n",
       "                                                        sw  \\\n",
       "0                                    of and the this <url>   \n",
       "1                                                   i have   \n",
       "2                          in the we are if we them in the   \n",
       "3                                    <user> i that i am so   \n",
       "4                      <user> <user> <user> i that to i am   \n",
       "...                                                    ...   \n",
       "1249995            i am so <user> to <user> <user> you are   \n",
       "1249996                                                  i   \n",
       "1249997  <user> same here but you over all the other to...   \n",
       "1249998                        off the for be by not <url>   \n",
       "1249999  <user> do not what has with me out or i don't ...   \n",
       "\n",
       "                                              without_punc          punc  \n",
       "0        vinco tresorpack 6 difficulti 10 10 object dis...       ( : ...  \n",
       "1                      glad dot tak tomorrow thank startho      ! ! # #   \n",
       "2           1 3 vs celtic regular season fuck play playoff          - =   \n",
       "3                             could actual kill girl sorri         ! ! !  \n",
       "4                             find veri hard believ afraid                \n",
       "...                                                    ...           ...  \n",
       "1249995                                  sorri guy ridicul   ! & & ! >:/  \n",
       "1249996                     cannot find food color anywher                \n",
       "1249997  tort wonder whi chose crime modul hahaa tri im...     ! ! ! ! ?  \n",
       "1249998  keyless entri remot fresh boat clicker 2005 bu...       - ( ...  \n",
       "1249999  yeap doctor know wrong laugh loud either migra...     . , . . .  \n",
       "\n",
       "[1250000 rows x 8 columns]"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "remove_punct_df(t_pos)\n",
    "remove_punct_df(t_neg)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "individual-reach",
   "metadata": {},
   "outputs": [],
   "source": [
    "count_pos_punct = t_pos.punc.str.split(expand = True).stack().value_counts()\n",
    "count_neg_punct = t_neg.punc.str.split(expand = True).stack().value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "naughty-updating",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "!       633765\n",
       ".       398593\n",
       ",       346296\n",
       "#       217015\n",
       "?       194681\n",
       "         ...  \n",
       "<->          1\n",
       ">:-{         1\n",
       "-'-          1\n",
       ";{           1\n",
       "]':          1\n",
       "Length: 210, dtype: int64"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "count_pos_punct"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "variable-capture",
   "metadata": {},
   "source": [
    "## 8. Most frequent words without special characters and without stopwords"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "oriented-analyst",
   "metadata": {},
   "outputs": [],
   "source": [
    "count_pos_no_punct_sw = t_pos.without_punc.str.split(expand = True).stack().value_counts()\n",
    "count_neg_no_punct_sw = t_neg.without_punc.str.split(expand = True).stack().value_counts()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "sustainable-creek",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "love             111247\n",
       "thank            108613\n",
       "follow           105317\n",
       "go                92135\n",
       "rt                87664\n",
       "                  ...  \n",
       "voxlap                1\n",
       "charlottetown         1\n",
       "collegewish           1\n",
       "scoreless             1\n",
       "ewallet               1\n",
       "Length: 224344, dtype: int64"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "count_pos_no_punct_sw"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "successful-eclipse",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "frame            102690\n",
       "go                98521\n",
       "want              74474\n",
       "get               63393\n",
       "2                 60049\n",
       "                  ...  \n",
       "notahapybunni         1\n",
       "60025                 1\n",
       "jpot                  1\n",
       "stamps                1\n",
       "ajeb                  1\n",
       "Length: 293889, dtype: int64"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "count_neg_no_punct_sw "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "competitive-republic",
   "metadata": {},
   "source": [
    "## 9. Plotting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "known-withdrawal",
   "metadata": {},
   "outputs": [],
   "source": [
    "value_neg = pd.DataFrame()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "metric-oriental",
   "metadata": {},
   "outputs": [],
   "source": [
    "neg_no_punct_sw = count_neg_no_punct_sw.to_frame(name = 'nbr').head(10)\n",
    "pos_no_punct_sw = count_pos_no_punct_sw.to_frame(name = 'nbr').head(10)\n",
    "\n",
    "pos_punct = count_pos_punct.to_frame(name = 'nbr').head(10)\n",
    "neg_punct = count_neg_punct.to_frame(name = 'nbr').head(10)\n",
    "\n",
    "pos_sw = count_pos_sw.to_frame(name = 'nbr').head(10)\n",
    "neg_sw = count_neg_sw.to_frame(name = 'nbr').head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "little-trademark",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>nbr</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>&lt;user&gt;</th>\n",
       "      <td>1027186</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>i</th>\n",
       "      <td>639229</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>you</th>\n",
       "      <td>528273</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>to</th>\n",
       "      <td>393353</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>the</th>\n",
       "      <td>339442</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>a</th>\n",
       "      <td>269448</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>is</th>\n",
       "      <td>238692</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>it</th>\n",
       "      <td>229659</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>and</th>\n",
       "      <td>226325</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>my</th>\n",
       "      <td>225578</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            nbr\n",
       "<user>  1027186\n",
       "i        639229\n",
       "you      528273\n",
       "to       393353\n",
       "the      339442\n",
       "a        269448\n",
       "is       238692\n",
       "it       229659\n",
       "and      226325\n",
       "my       225578"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pos_sw"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "dense-burner",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.25, -0.06, 'Most frequently appearing strings')"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 936x720 with 6 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "to_plot = [pos_no_punct_sw, pos_sw, pos_punct, neg_no_punct_sw, neg_sw, neg_punct]\n",
    "\n",
    "#to_plot = {pos_no_punct_sw : 'Words without punctuation nor stop wods', pos_sw :}\n",
    "\n",
    "fig, ax = plt.subplots(3, 2, figsize = (13, 10), sharey = 'row', constrained_layout =True)\n",
    "\n",
    "x = np.arange(10)\n",
    "\n",
    "for i in range(2):\n",
    "    for j in range(3):\n",
    "        names = to_plot[i*3+j].index\n",
    "        values = to_plot[i*3+j].nbr\n",
    "        ax[j, i].bar(names, values)\n",
    "        ax[j, i].tick_params(axis = 'y', labelsize =15 )      \n",
    "        if j==2:\n",
    "            ax[j, i].tick_params(axis = 'x', labelsize=20)\n",
    "        else :\n",
    "            ax[j, i].tick_params(axis = 'x', labelsize =15 )\n",
    "\n",
    "#plt.rc('xtick', labelsize = 30)\n",
    "fig.suptitle(\"Top-10 occurring strings for each category\", fontsize = 40)\n",
    "fig.text(-0.05,0.3, \"Number of occurences\", rotation = 90, fontsize=30)\n",
    "fig.text(-0.02,0.75, \"Tokens\", rotation = 90, fontsize=20)\n",
    "fig.text(-0.02,0.40, \"Stop words\", rotation = 90, fontsize=20)\n",
    "fig.text(-0.02,0.10, \"Punctuation\", rotation = 90, fontsize=20)\n",
    "fig.text(0.15,-0.02, \"Positive tweets\", fontsize=20)\n",
    "fig.text(0.65, -0.02, \"Negative tweets\", fontsize = 20)\n",
    "fig.text(0.25, -0.06, \"Most frequently appearing strings\", fontsize=30)\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "caring-programming",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}