NLP_Text_processing_and_Intro.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "857adfd0",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package punkt to /Users/akshays/nltk_data...\n",
      "[nltk_data]   Package punkt is already up-to-date!\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import os\n",
    "import re \n",
    "import random\n",
    "import string      # for string operations    \n",
    "import pandas as pd\n",
    "import numpy as np     \n",
    "import matplotlib.pyplot as plt \n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "\n",
    "import nltk\n",
    "nltk.download('punkt')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "ec48aa86",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Collecting gdown\n",
      "  Downloading gdown-4.7.1-py3-none-any.whl (15 kB)\n",
      "Requirement already satisfied: filelock in /Users/akshays/opt/anaconda3/envs/nlp_env/lib/python3.11/site-packages (from gdown) (3.9.0)\n",
      "Requirement already satisfied: requests[socks] in /Users/akshays/opt/anaconda3/envs/nlp_env/lib/python3.11/site-packages (from gdown) (2.29.0)\n",
      "Requirement already satisfied: six in /Users/akshays/opt/anaconda3/envs/nlp_env/lib/python3.11/site-packages (from gdown) (1.16.0)\n",
      "Requirement already satisfied: tqdm in /Users/akshays/opt/anaconda3/envs/nlp_env/lib/python3.11/site-packages (from gdown) (4.65.0)\n",
      "Requirement already satisfied: beautifulsoup4 in /Users/akshays/opt/anaconda3/envs/nlp_env/lib/python3.11/site-packages (from gdown) (4.12.2)\n",
      "Requirement already satisfied: soupsieve>1.2 in /Users/akshays/opt/anaconda3/envs/nlp_env/lib/python3.11/site-packages (from beautifulsoup4->gdown) (2.4)\n",
      "Requirement already satisfied: charset-normalizer<4,>=2 in /Users/akshays/opt/anaconda3/envs/nlp_env/lib/python3.11/site-packages (from requests[socks]->gdown) (2.0.4)\n",
      "Requirement already satisfied: idna<4,>=2.5 in /Users/akshays/opt/anaconda3/envs/nlp_env/lib/python3.11/site-packages (from requests[socks]->gdown) (3.4)\n",
      "Requirement already satisfied: urllib3<1.27,>=1.21.1 in /Users/akshays/opt/anaconda3/envs/nlp_env/lib/python3.11/site-packages (from requests[socks]->gdown) (1.26.16)\n",
      "Requirement already satisfied: certifi>=2017.4.17 in /Users/akshays/opt/anaconda3/envs/nlp_env/lib/python3.11/site-packages (from requests[socks]->gdown) (2023.5.7)\n",
      "Requirement already satisfied: PySocks!=1.5.7,>=1.5.6 in /Users/akshays/opt/anaconda3/envs/nlp_env/lib/python3.11/site-packages (from requests[socks]->gdown) (1.7.1)\n",
      "Installing collected packages: gdown\n",
      "Successfully installed gdown-4.7.1\n",
      "Note: you may need to restart the kernel to use updated packages.\n"
     ]
    }
   ],
   "source": [
    "#pip install gdown"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d7c6dd16",
   "metadata": {},
   "source": [
    "# Sentiment analysis- Twiter(tweets)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "b83c9c7c",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading...\n",
      "From: https://drive.google.com/uc?id=11LWo5nagSmC72hOJM9lm2l05sdQxt4rf\n",
      "To: /Users/akshays/Documents/Personal Projects/NLP practivce/NLP_practice/corona_tweets.csv\n",
      "100%|██████████████████████████████████████| 1.00M/1.00M [00:00<00:00, 3.22MB/s]\n"
     ]
    }
   ],
   "source": [
    "!gdown 11LWo5nagSmC72hOJM9lm2l05sdQxt4rf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "d85c4bb2",
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('corona_tweets.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "2d4187ef",
   "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>UserName</th>\n",
       "      <th>ScreenName</th>\n",
       "      <th>Location</th>\n",
       "      <th>TweetAt</th>\n",
       "      <th>OriginalTweet</th>\n",
       "      <th>Sentiment</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>44953</td>\n",
       "      <td>NYC</td>\n",
       "      <td>02-03-2020</td>\n",
       "      <td>TRENDING: New Yorkers encounter empty supermar...</td>\n",
       "      <td>Extremely Negative</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>44954</td>\n",
       "      <td>Seattle, WA</td>\n",
       "      <td>02-03-2020</td>\n",
       "      <td>When I couldn't find hand sanitizer at Fred Me...</td>\n",
       "      <td>Positive</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>44955</td>\n",
       "      <td>NaN</td>\n",
       "      <td>02-03-2020</td>\n",
       "      <td>Find out how you can protect yourself and love...</td>\n",
       "      <td>Extremely Positive</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>44956</td>\n",
       "      <td>Chicagoland</td>\n",
       "      <td>02-03-2020</td>\n",
       "      <td>#Panic buying hits #NewYork City as anxious sh...</td>\n",
       "      <td>Negative</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>44957</td>\n",
       "      <td>Melbourne, Victoria</td>\n",
       "      <td>03-03-2020</td>\n",
       "      <td>#toiletpaper #dunnypaper #coronavirus #coronav...</td>\n",
       "      <td>Neutral</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>3793</th>\n",
       "      <td>3794</td>\n",
       "      <td>48746</td>\n",
       "      <td>Israel ??</td>\n",
       "      <td>16-03-2020</td>\n",
       "      <td>Meanwhile In A Supermarket in Israel -- People...</td>\n",
       "      <td>Positive</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3794</th>\n",
       "      <td>3795</td>\n",
       "      <td>48747</td>\n",
       "      <td>Farmington, NM</td>\n",
       "      <td>16-03-2020</td>\n",
       "      <td>Did you panic buy a lot of non-perishable item...</td>\n",
       "      <td>Negative</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3795</th>\n",
       "      <td>3796</td>\n",
       "      <td>48748</td>\n",
       "      <td>Haverford, PA</td>\n",
       "      <td>16-03-2020</td>\n",
       "      <td>Asst Prof of Economics @cconces was on @NBCPhi...</td>\n",
       "      <td>Neutral</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3796</th>\n",
       "      <td>3797</td>\n",
       "      <td>48749</td>\n",
       "      <td>NaN</td>\n",
       "      <td>16-03-2020</td>\n",
       "      <td>Gov need to do somethings instead of biar je r...</td>\n",
       "      <td>Extremely Negative</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3797</th>\n",
       "      <td>3798</td>\n",
       "      <td>48750</td>\n",
       "      <td>Arlington, Virginia</td>\n",
       "      <td>16-03-2020</td>\n",
       "      <td>I and @ForestandPaper members are committed to...</td>\n",
       "      <td>Extremely Positive</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3798 rows × 6 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      UserName  ScreenName             Location     TweetAt  \\\n",
       "0            1       44953                  NYC  02-03-2020   \n",
       "1            2       44954          Seattle, WA  02-03-2020   \n",
       "2            3       44955                  NaN  02-03-2020   \n",
       "3            4       44956          Chicagoland  02-03-2020   \n",
       "4            5       44957  Melbourne, Victoria  03-03-2020   \n",
       "...        ...         ...                  ...         ...   \n",
       "3793      3794       48746            Israel ??  16-03-2020   \n",
       "3794      3795       48747       Farmington, NM  16-03-2020   \n",
       "3795      3796       48748        Haverford, PA  16-03-2020   \n",
       "3796      3797       48749                  NaN  16-03-2020   \n",
       "3797      3798       48750  Arlington, Virginia  16-03-2020   \n",
       "\n",
       "                                          OriginalTweet           Sentiment  \n",
       "0     TRENDING: New Yorkers encounter empty supermar...  Extremely Negative  \n",
       "1     When I couldn't find hand sanitizer at Fred Me...            Positive  \n",
       "2     Find out how you can protect yourself and love...  Extremely Positive  \n",
       "3     #Panic buying hits #NewYork City as anxious sh...            Negative  \n",
       "4     #toiletpaper #dunnypaper #coronavirus #coronav...             Neutral  \n",
       "...                                                 ...                 ...  \n",
       "3793  Meanwhile In A Supermarket in Israel -- People...            Positive  \n",
       "3794  Did you panic buy a lot of non-perishable item...            Negative  \n",
       "3795  Asst Prof of Economics @cconces was on @NBCPhi...             Neutral  \n",
       "3796  Gov need to do somethings instead of biar je r...  Extremely Negative  \n",
       "3797  I and @ForestandPaper members are committed to...  Extremely Positive  \n",
       "\n",
       "[3798 rows x 6 columns]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "b3a0ed34",
   "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>UserName</th>\n",
       "      <th>ScreenName</th>\n",
       "      <th>Location</th>\n",
       "      <th>TweetAt</th>\n",
       "      <th>OriginalTweet</th>\n",
       "      <th>Sentiment</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>3798.000000</td>\n",
       "      <td>3798.000000</td>\n",
       "      <td>2964</td>\n",
       "      <td>3798</td>\n",
       "      <td>3798</td>\n",
       "      <td>3798</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>unique</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1717</td>\n",
       "      <td>15</td>\n",
       "      <td>3798</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>top</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>United States</td>\n",
       "      <td>13-03-2020</td>\n",
       "      <td>TRENDING: New Yorkers encounter empty supermar...</td>\n",
       "      <td>Negative</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>freq</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>75</td>\n",
       "      <td>1233</td>\n",
       "      <td>1</td>\n",
       "      <td>1041</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>1899.500000</td>\n",
       "      <td>46851.500000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>1096.532489</td>\n",
       "      <td>1096.532489</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>44953.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>950.250000</td>\n",
       "      <td>45902.250000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>1899.500000</td>\n",
       "      <td>46851.500000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>2848.750000</td>\n",
       "      <td>47800.750000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>3798.000000</td>\n",
       "      <td>48750.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           UserName    ScreenName       Location     TweetAt  \\\n",
       "count   3798.000000   3798.000000           2964        3798   \n",
       "unique          NaN           NaN           1717          15   \n",
       "top             NaN           NaN  United States  13-03-2020   \n",
       "freq            NaN           NaN             75        1233   \n",
       "mean    1899.500000  46851.500000            NaN         NaN   \n",
       "std     1096.532489   1096.532489            NaN         NaN   \n",
       "min        1.000000  44953.000000            NaN         NaN   \n",
       "25%      950.250000  45902.250000            NaN         NaN   \n",
       "50%     1899.500000  46851.500000            NaN         NaN   \n",
       "75%     2848.750000  47800.750000            NaN         NaN   \n",
       "max     3798.000000  48750.000000            NaN         NaN   \n",
       "\n",
       "                                            OriginalTweet Sentiment  \n",
       "count                                                3798      3798  \n",
       "unique                                               3798         5  \n",
       "top     TRENDING: New Yorkers encounter empty supermar...  Negative  \n",
       "freq                                                    1      1041  \n",
       "mean                                                  NaN       NaN  \n",
       "std                                                   NaN       NaN  \n",
       "min                                                   NaN       NaN  \n",
       "25%                                                   NaN       NaN  \n",
       "50%                                                   NaN       NaN  \n",
       "75%                                                   NaN       NaN  \n",
       "max                                                   NaN       NaN  "
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.describe(include='all')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "6fa45daa",
   "metadata": {},
   "outputs": [],
   "source": [
    "df['Sentiment'] = df['Sentiment'].replace('Extremely Positive','Positive' )\n",
    "df['Sentiment'] = df['Sentiment'].replace('Extremely Negative','Negative' )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "d3b92078",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Negative    1633\n",
       "Positive    1546\n",
       "Neutral      619\n",
       "Name: Sentiment, dtype: int64"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['Sentiment'].value_countsnts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "79ff3aa1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Size of training dataset:  1993\n",
      "Size of testing dataset:  500\n",
      "\u001b[92m\n",
      "positive in green: \n",
      "\u001b[92mComing back from the grocery store like... #CoronavirusPandemic #Covid_19 #lockdown https://t.co/9AW8hDr8Mo\n",
      "\u001b[91m\n",
      "negative in red: \n",
      "\u001b[91mThe economic effect of Covid-19 will likely reduce global demand for oil and will effect other sectors. Saudi Arabia has forced oil prices down to $30 a barrel as well. \r",
      "\r\n",
      "\r",
      "\r\n",
      "#ahg #finance x #financeservices #coronavirus #oilvscoronavirus #gulfcountires #economy #globaleffect https://t.co/cU77CasWG3\n"
     ]
    }
   ],
   "source": [
    "# Separating positive_tweets and negative_tweets\n",
    "all_positive_tweets = list(df[df['Sentiment'] == 'Positive']['OriginalTweet'].astype(str)) # Making equal positive and negative tweet counts\n",
    "all_negative_tweets = list(df[df['Sentiment'] == 'Negative']['OriginalTweet'].astype(str))[:947]\n",
    "\n",
    "# Splitting training and testing set. 80/20 split\n",
    "positive_train, positive_test = train_test_split(all_positive_tweets, test_size=0.2, random_state=42)\n",
    "negative_train, negative_test = train_test_split(all_negative_tweets, test_size=0.2, random_state=42)\n",
    "\n",
    "print(\"Size of training dataset: \",len(positive_train) + len(negative_train))\n",
    "print(\"Size of testing dataset: \", len(positive_test)  + len(negative_test))\n",
    "\n",
    "# print positive in green\n",
    "print('\\033[92m' + '\\npositive in green: ')\n",
    "print('\\033[92m' + all_positive_tweets[random.randint(0,947)])\n",
    "print('\\033[91m' + '\\nnegative in red: ')\n",
    "# print negative in red\n",
    "print('\\033[91m' + all_negative_tweets[random.randint(0,947)])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b6c5bd53",
   "metadata": {},
   "source": [
    "# Tokenization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "c0925c55",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using Split : ['We', '(@hamiltonpoverty)', 'are', 'calling', 'on', 'Canadian', 'Federal', 'Government', '??', 'to', 'provide', '#BasicIncome', \"'Well-being\", 'and', 'Stimulus', \"Payments'\", 'during', 'current', 'COVID-19', 'situation.', 'A', 'basic', 'income', '(even', 'interim)', 'could', 'help', 'keep', 'people', 'healthy,', 'then', 'revive', 'consumer', 'confidence', 'and', 'spending.', '#CdnPoli', 'https://t.co/wRmkWLhJku']\n"
     ]
    }
   ],
   "source": [
    "# Split on spaces\n",
    "tweet = all_positive_tweets[43] \n",
    "word = tweet.split() # space tokenizer\n",
    "print('Using Split :',word)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6a828c83",
   "metadata": {},
   "source": [
    "### Python already has modules for collecting, handling, and processing text data.\n",
    "\n",
    "\n",
    "> **Using Natural Language Toolkit (NLTK) package.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "41392655",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using NLTK : ['We', '(', '@', 'hamiltonpoverty', ')', 'are', 'calling', 'on', 'Canadian', 'Federal', 'Government', '?', '?', 'to', 'provide', '#', 'BasicIncome', \"'Well-being\", 'and', 'Stimulus', 'Payments', \"'\", 'during', 'current', 'COVID-19', 'situation', '.', 'A', 'basic', 'income', '(', 'even', 'interim', ')', 'could', 'help', 'keep', 'people', 'healthy', ',', 'then', 'revive', 'consumer', 'confidence', 'and', 'spending', '.', '#', 'CdnPoli', 'https', ':', '//t.co/wRmkWLhJku']\n"
     ]
    }
   ],
   "source": [
    "from nltk.tokenize import word_tokenize\n",
    "\n",
    "words = word_tokenize(tweet)\n",
    "print('Using NLTK :',words)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "a7be2a60",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Python library for NLP\n",
    "from nltk.corpus import twitter_samples    # sample Twitter dataset from NLTK\n",
    "from nltk.corpus import stopwords          # module for stop words that come with NLTK\n",
    "from nltk.stem import PorterStemmer        # module for stemming\n",
    "from nltk.tokenize import TweetTokenizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "f2df20d4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\"We (@hamiltonpoverty) are calling on Canadian Federal Government ?? to provide #BasicIncome 'Well-being and Stimulus Payments' during current COVID-19 situation. A basic income (even interim) could help keep people healthy, then revive consumer confidence and spending. #CdnPoli https://t.co/wRmkWLhJku\""
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tweet = all_positive_tweets[43] \n",
    "tweet"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "1d4bfc70",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[92mWe (@hamiltonpoverty) are calling on Canadian Federal Government ?? to provide #BasicIncome 'Well-being and Stimulus Payments' during current COVID-19 situation. A basic income (even interim) could help keep people healthy, then revive consumer confidence and spending. #CdnPoli https://t.co/wRmkWLhJku\n",
      "\u001b[94m\n",
      "We (@hamiltonpoverty) are calling on Canadian Federal Government ?? to provide BasicIncome 'Well-being and Stimulus Payments' during current COVID-19 situation. A basic income (even interim) could help keep people healthy, then revive consumer confidence and spending. CdnPoli \n"
     ]
    }
   ],
   "source": [
    "# Removing hyperlinks and hashtags\n",
    "\n",
    "print('\\033[92m' + tweet)\n",
    "print('\\033[94m')\n",
    "\n",
    "# remove old style retweet text \"RT\"\n",
    "tweet2 = re.sub(r'^RT[\\s]+', '', tweet)\n",
    "\n",
    "# remove hyperlinks\n",
    "tweet2 = re.sub(r'https?://[^\\s\\n\\r]+', '', tweet2)\n",
    "\n",
    "# remove hashtags\n",
    "# only removing the hash # sign from the word\n",
    "tweet2 = re.sub(r'#', '', tweet2)\n",
    "print(tweet2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "e61eca5c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "\u001b[92mWe (@hamiltonpoverty) are calling on Canadian Federal Government ?? to provide BasicIncome 'Well-being and Stimulus Payments' during current COVID-19 situation. A basic income (even interim) could help keep people healthy, then revive consumer confidence and spending. CdnPoli \n",
      "\u001b[94m\n",
      "\n",
      "Tokenized string:\n",
      "['we', '(', ')', 'are', 'calling', 'on', 'canadian', 'federal', 'government', '?', '?', 'to', 'provide', 'basicincome', \"'\", 'well-being', 'and', 'stimulus', 'payments', \"'\", 'during', 'current', 'covid', '-', '19', 'situation', '.', 'a', 'basic', 'income', '(', 'even', 'interim', ')', 'could', 'help', 'keep', 'people', 'healthy', ',', 'then', 'revive', 'consumer', 'confidence', 'and', 'spending', '.', 'cdnpoli']\n"
     ]
    }
   ],
   "source": [
    "# Tokenizing\n",
    "\n",
    "print()\n",
    "print('\\033[92m' + tweet2)\n",
    "print('\\033[94m')\n",
    "\n",
    "# instantiate tokenizer class\n",
    "tokenizer = TweetTokenizer(preserve_case=False, strip_handles=True,\n",
    "                               reduce_len=True)\n",
    "\n",
    "# tokenize tweets\n",
    "tweet_tokens = tokenizer.tokenize(tweet2)\n",
    "\n",
    "print()\n",
    "print('Tokenized string:')\n",
    "print(tweet_tokens)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "c559ca4a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Stop words\n",
      "\n",
      "['i', 'me', 'my', 'myself', 'we', 'our', 'ours', 'ourselves', 'you', \"you're\", \"you've\", \"you'll\", \"you'd\", 'your', 'yours', 'yourself', 'yourselves', 'he', 'him', 'his', 'himself', 'she', \"she's\", 'her', 'hers', 'herself', 'it', \"it's\", 'its', 'itself', 'they', 'them', 'their', 'theirs', 'themselves', 'what', 'which', 'who', 'whom', 'this', 'that', \"that'll\", 'these', 'those', 'am', 'is', 'are', 'was', 'were', 'be', 'been', 'being', 'have', 'has', 'had', 'having', 'do', 'does', 'did', 'doing', 'a', 'an', 'the', 'and', 'but', 'if', 'or', 'because', 'as', 'until', 'while', 'of', 'at', 'by', 'for', 'with', 'about', 'against', 'between', 'into', 'through', 'during', 'before', 'after', 'above', 'below', 'to', 'from', 'up', 'down', 'in', 'out', 'on', 'off', 'over', 'under', 'again', 'further', 'then', 'once', 'here', 'there', 'when', 'where', 'why', 'how', 'all', 'any', 'both', 'each', 'few', 'more', 'most', 'other', 'some', 'such', 'no', 'nor', 'not', 'only', 'own', 'same', 'so', 'than', 'too', 'very', 's', 't', 'can', 'will', 'just', 'don', \"don't\", 'should', \"should've\", 'now', 'd', 'll', 'm', 'o', 're', 've', 'y', 'ain', 'aren', \"aren't\", 'couldn', \"couldn't\", 'didn', \"didn't\", 'doesn', \"doesn't\", 'hadn', \"hadn't\", 'hasn', \"hasn't\", 'haven', \"haven't\", 'isn', \"isn't\", 'ma', 'mightn', \"mightn't\", 'mustn', \"mustn't\", 'needn', \"needn't\", 'shan', \"shan't\", 'shouldn', \"shouldn't\", 'wasn', \"wasn't\", 'weren', \"weren't\", 'won', \"won't\", 'wouldn', \"wouldn't\"]\n",
      "\n",
      "Punctuation\n",
      "\n",
      "!\"#$%&'()*+,-./:;<=>?@[\\]^_`{|}~\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package stopwords to\n",
      "[nltk_data]     /Users/akshays/nltk_data...\n",
      "[nltk_data]   Package stopwords is already up-to-date!\n"
     ]
    }
   ],
   "source": [
    "# Import the english stop words list from NLTK\n",
    "nltk.download(\"stopwords\")\n",
    "stopwords_english = stopwords.words('english') \n",
    "\n",
    "print('Stop words\\n')\n",
    "print(stopwords_english)\n",
    "\n",
    "print('\\nPunctuation\\n')\n",
    "print(string.punctuation)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "d5717e10",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "\u001b[92m\n",
      "['we', '(', ')', 'are', 'calling', 'on', 'canadian', 'federal', 'government', '?', '?', 'to', 'provide', 'basicincome', \"'\", 'well-being', 'and', 'stimulus', 'payments', \"'\", 'during', 'current', 'covid', '-', '19', 'situation', '.', 'a', 'basic', 'income', '(', 'even', 'interim', ')', 'could', 'help', 'keep', 'people', 'healthy', ',', 'then', 'revive', 'consumer', 'confidence', 'and', 'spending', '.', 'cdnpoli']\n",
      "\u001b[94m\n",
      "removed stop words and punctuation:\n",
      "['calling', 'canadian', 'federal', 'government', 'provide', 'basicincome', 'well-being', 'stimulus', 'payments', 'current', 'covid', '19', 'situation', 'basic', 'income', 'even', 'interim', 'could', 'help', 'keep', 'people', 'healthy', 'revive', 'consumer', 'confidence', 'spending', 'cdnpoli']\n"
     ]
    }
   ],
   "source": [
    "print()\n",
    "print('\\033[92m')\n",
    "print(tweet_tokens)\n",
    "print('\\033[94m')\n",
    "\n",
    "tweets_clean = []\n",
    "\n",
    "for word in tweet_tokens: # Go through every word in your tokens list\n",
    "    if (word not in stopwords_english and  # remove stopwords\n",
    "        word not in string.punctuation):  # remove punctuation\n",
    "        tweets_clean.append(word)\n",
    "\n",
    "print('removed stop words and punctuation:')\n",
    "print(tweets_clean)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "b54f8577",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "\u001b[92m\n",
      "['calling', 'canadian', 'federal', 'government', 'provide', 'basicincome', 'well-being', 'stimulus', 'payments', 'current', 'covid', '19', 'situation', 'basic', 'income', 'even', 'interim', 'could', 'help', 'keep', 'people', 'healthy', 'revive', 'consumer', 'confidence', 'spending', 'cdnpoli']\n",
      "\u001b[94m\n",
      "stemmed words:\n",
      "['call', 'canadian', 'feder', 'govern', 'provid', 'basicincom', 'well-b', 'stimulu', 'payment', 'current', 'covid', '19', 'situat', 'basic', 'incom', 'even', 'interim', 'could', 'help', 'keep', 'peopl', 'healthi', 'reviv', 'consum', 'confid', 'spend', 'cdnpoli']\n"
     ]
    }
   ],
   "source": [
    "print()\n",
    "print('\\033[92m')\n",
    "print(tweets_clean)\n",
    "print('\\033[94m')\n",
    "\n",
    "# Instantiate stemming class\n",
    "stemmer = PorterStemmer() \n",
    "\n",
    "# Create an empty list to store the stems\n",
    "tweets_stem = [] \n",
    "\n",
    "for word in tweets_clean:\n",
    "    stem_word = stemmer.stem(word)  # stemming word\n",
    "    tweets_stem.append(stem_word)  # append to the list\n",
    "\n",
    "print('stemmed words:')\n",
    "print(tweets_stem)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "9c806463",
   "metadata": {},
   "outputs": [],
   "source": [
    "def process_tweet(tweet):\n",
    "    stemmer = PorterStemmer()\n",
    "    stopwords_english = stopwords.words('english')\n",
    "    # remove stock market tickers like $GE\n",
    "    tweet = re.sub(r'\\$\\w*', '', tweet)\n",
    "    # remove old style retweet text \"RT\"\n",
    "    tweet = re.sub(r'^RT[\\s]+', '', tweet)\n",
    "    # remove hyperlinks    \n",
    "    tweet = re.sub(r'https?://[^\\s\\n\\r]+', '', tweet)\n",
    "    # remove hashtags\n",
    "    # only removing the hash # sign from the word\n",
    "    tweet = re.sub(r'#', '', tweet)\n",
    "    # tokenize tweets\n",
    "    tokenizer = TweetTokenizer(preserve_case=False, strip_handles=True,\n",
    "                               reduce_len=True)\n",
    "    tweet_tokens = tokenizer.tokenize(tweet)\n",
    "\n",
    "    tweets_clean = []\n",
    "    for word in tweet_tokens:\n",
    "        if (word not in stopwords_english and  # remove stopwords\n",
    "                word not in string.punctuation):  # remove punctuation\n",
    "            # tweets_clean.append(word)\n",
    "            stem_word = stemmer.stem(word)  # stemming word\n",
    "            tweets_clean.append(stem_word)\n",
    "\n",
    "    return tweets_clean"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "4f00edab",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "\u001b[92m\n",
      "We (@hamiltonpoverty) are calling on Canadian Federal Government ?? to provide #BasicIncome 'Well-being and Stimulus Payments' during current COVID-19 situation. A basic income (even interim) could help keep people healthy, then revive consumer confidence and spending. #CdnPoli https://t.co/wRmkWLhJku\n",
      "\u001b[94m\n",
      "preprocessed tweet:\n",
      "['call', 'canadian', 'feder', 'govern', 'provid', 'basicincom', 'well-b', 'stimulu', 'payment', 'current', 'covid', '19', 'situat', 'basic', 'incom', 'even', 'interim', 'could', 'help', 'keep', 'peopl', 'healthi', 'reviv', 'consum', 'confid', 'spend', 'cdnpoli']\n"
     ]
    }
   ],
   "source": [
    "# choose the same tweet\n",
    "tweet = all_positive_tweets[43] \n",
    "\n",
    "print()\n",
    "print('\\033[92m')\n",
    "print(tweet)\n",
    "print('\\033[94m')\n",
    "\n",
    "# call the imported function\n",
    "tweets_stem = process_tweet(tweet); # Preprocess a given tweet\n",
    "\n",
    "print('preprocessed tweet:')\n",
    "print(tweets_stem) # Print the result"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "96084f3e",
   "metadata": {},
   "source": [
    "# BOW/ Count BOWs\n",
    "Now we can build a BOWs/ Count BOWs, in which we will create matrix of 1s and 0s by using our vocabulary for each tweet!--- check for dimensionality and sparseness of the vector"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "9d9c4389",
   "metadata": {},
   "outputs": [],
   "source": [
    "def build_freqs(tweets, ys):\n",
    "    \"\"\"\n",
    "        tweets: a list of tweets\n",
    "        ys: an m x 1 array with the sentiment label of each tweet\n",
    "            (either 0 or 1)\n",
    "        freqs: a dictionary mapping each (word, sentiment) pair (=key) to its frequency (=value)\n",
    "    \"\"\"\n",
    "    # Convert np array to list since zip needs an iterable.\n",
    "    # The squeeze is necessary or the list ends up with one element.\n",
    "    # Also note that this is just a NOP if ys is already a list.\n",
    "    yslist = np.squeeze(ys).tolist()\n",
    "\n",
    "    # Start with an empty dictionary and populate it by looping over all tweets\n",
    "    # and over all processed words in each tweet.\n",
    "    freqs = {}\n",
    "    for y, tweet in zip(yslist, tweets):\n",
    "        for word in process_tweet(tweet):\n",
    "            pair = (word, y)\n",
    "            if pair in freqs:\n",
    "                freqs[pair] += 1\n",
    "            else:\n",
    "                freqs[pair] = 1\n",
    "\n",
    "    return freqs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "1ffb7266",
   "metadata": {},
   "outputs": [],
   "source": [
    "training_tweets = positive_train + negative_train\n",
    "# make a numpy array representing labels of the tweets\n",
    "labels = np.append(np.ones((len(positive_train))), np.zeros((len(negative_train))))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "cad8c923",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "type(freqs) = <class 'dict'>\n",
      "len(freqs) = 8007\n"
     ]
    }
   ],
   "source": [
    "# create frequency dictionary\n",
    "freqs = build_freqs(training_tweets, labels)\n",
    "\n",
    "# check data type\n",
    "print(f'type(freqs) = {type(freqs)}')\n",
    "\n",
    "# check length of the dictionary\n",
    "print(f'len(freqs) = {len(freqs)}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "d7bd341c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[['void', 1, 0],\n",
       " ['commun', 31, 9],\n",
       " ['spirit', 2, 1],\n",
       " ['stop', 22, 36],\n",
       " ['merri', 0, 0],\n",
       " ['nice', 15, 4],\n",
       " ['good', 96, 25],\n",
       " ['bad', 3, 16],\n",
       " ['sad', 0, 7],\n",
       " ['mad', 4, 10],\n",
       " ['best', 35, 3],\n",
       " ['pretti', 12, 1],\n",
       " [':)', 4, 1],\n",
       " [':(', 3, 0],\n",
       " ['song', 2, 0],\n",
       " ['idea', 11, 6],\n",
       " ['power', 11, 1],\n",
       " ['play', 17, 1]]"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# select some words to appear in the report.\n",
    "keys = ['void', 'commun', 'spirit', 'stop', 'merri', 'nice', 'good', 'bad', 'sad', 'mad', 'best', 'pretti',\n",
    "        ':)', ':(', 'song', 'idea', 'power', 'play']\n",
    "\n",
    "# list representing our table of word counts.\n",
    "# each element consist of a sublist with this pattern: [<word>, <positive_count>, <negative_count>]\n",
    "data = []\n",
    "\n",
    "# loop through our selected words\n",
    "for word in keys:\n",
    "    \n",
    "    # initialize positive and negative counts\n",
    "    pos = 0\n",
    "    neg = 0\n",
    "    \n",
    "    # retrieve number of positive counts\n",
    "    if (word, 1) in freqs:\n",
    "        pos = freqs[(word, 1)]\n",
    "        \n",
    "    # retrieve number of negative counts\n",
    "    if (word, 0) in freqs:\n",
    "        neg = freqs[(word, 0)]\n",
    "        \n",
    "    # append the word counts to the table\n",
    "    data.append([word, pos, neg])\n",
    "    \n",
    "data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "3f97480f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x800 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize = (8, 8))\n",
    "\n",
    "# convert positive raw counts to logarithmic scale. we add 1 to avoid log(0)\n",
    "x = np.log([x[1] + 1 for x in data])  \n",
    "\n",
    "# do the same for the negative counts\n",
    "y = np.log([x[2] + 1 for x in data]) \n",
    "\n",
    "# Plot a dot for each pair of words\n",
    "ax.scatter(x, y)  \n",
    "\n",
    "# assign axis labels\n",
    "plt.xlabel(\"Log Positive count\")\n",
    "plt.ylabel(\"Log Negative count\")\n",
    "\n",
    "# Add the word as the label at the same position as you added the points just before\n",
    "for i in range(0, len(data)):\n",
    "    ax.annotate(data[i][0], (x[i], y[i]), fontsize=12)\n",
    "\n",
    "ax.plot([0, 9], [0, 9], color = 'brown') # Plot the red line that divides the 2 areas.\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "c8fd8e2d",
   "metadata": {},
   "outputs": [],
   "source": [
    "def extract_features(tweet, freqs):\n",
    "    '''\n",
    "        tweet: a list of words for one tweet\n",
    "        freqs: a dictionary corresponding to the frequencies of each tuple (word, label)\n",
    "        x: a feature vector of dimension (1,3)\n",
    "    '''\n",
    "    # process_tweet tokenizes, stems, and removes stopwords\n",
    "    word_l = process_tweet(tweet)\n",
    "    # 3 elements in the form of a 1 x 3 vector\n",
    "    x = np.zeros((1, 3)) \n",
    "    \n",
    "    #bias term is set to 1\n",
    "    x[0,0] = 1 \n",
    "        \n",
    "    # loop through each word in the list of words\n",
    "    for word in word_l:\n",
    "        \n",
    "        # increment the word count for the positive label 1\n",
    "        x[0,1] += freqs.get((word, 1.0),0)\n",
    "        \n",
    "        # increment the word count for the negative label 0\n",
    "        x[0,2] += freqs.get((word, 0.0),0)\n",
    "        \n",
    "    assert(x.shape == (1, 3))\n",
    "    return x"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "da737ffd",
   "metadata": {},
   "source": [
    "# Extract feature for a tweet"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "1a93abf2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[1.000e+00 2.418e+03 1.642e+03]]\n"
     ]
    }
   ],
   "source": [
    "train_x = training_tweets\n",
    "tmp1 = extract_features(train_x[0], freqs)\n",
    "print(tmp1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "ec4017a2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[ 1. 39.  3.]]\n"
     ]
    }
   ],
   "source": [
    "tmp2 = extract_features('This batch is the best batch', freqs)\n",
    "print(tmp2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "68e909a4",
   "metadata": {},
   "outputs": [],
   "source": [
    "# collect the features 'x' and stack them into a matrix 'X'\n",
    "X = np.zeros((len(train_x), 3))\n",
    "for i in range(len(train_x)):\n",
    "    X[i, :]= extract_features(train_x[i], freqs)\n",
    "\n",
    "# training labels corresponding to X\n",
    "train_y = np.append(np.ones((len(positive_train), 1)), np.zeros((len(negative_train), 1)), axis=0)\n",
    "Y = np.ravel(train_y,order='C')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a70a260a",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "ee81050d",
   "metadata": {},
   "source": [
    "# Most basic- Logistic regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "b9efde9c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy:  0.703963873557451\n"
     ]
    }
   ],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "clf = LogisticRegression().fit(X, Y)\n",
    "\n",
    "print(\"Training Accuracy: \",clf.score(X, Y))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "61ef18ae",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Testing Accuracy:  0.648\n"
     ]
    }
   ],
   "source": [
    "# Testing\n",
    "testing_tweets = positive_test + negative_test\n",
    "\n",
    "test_X = np.zeros((len(testing_tweets), 3))\n",
    "for i in range(len(testing_tweets)):\n",
    "    test_X[i, :]= extract_features(testing_tweets[i], freqs)\n",
    "\n",
    "# training labels corresponding to X\n",
    "test_y = np.append(np.ones((len(positive_test), 1)), np.zeros((len(negative_test), 1)), axis=0)\n",
    "test_Y = np.ravel(test_y,order='C')\n",
    "\n",
    "print(\"Testing Accuracy: \",clf.score(test_X, test_Y))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "6fcbcc83",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[92mThank God coronovirus is over -> 1.000000\n",
      "\u001b[91mThis is seriously ridiculous. Stop hoarding -> 1.000000\n",
      "\u001b[92mGovernment efforts are great. -> 1.000000\n",
      "\u001b[93mNonsense -> 1.000000\n",
      "\u001b[91mPanic-buying is pushing up prices.#coronavirus -> 1.000000\n",
      "\u001b[92mTHANK YOU to make hand sanitizer in distillery amid outbreak -> 1.000000\n"
     ]
    }
   ],
   "source": [
    "def predict_custom_tweet(tweet, freqs):\n",
    "    x = extract_features(tweet,freqs)\n",
    "    y_pred = clf.predict(x)\n",
    "    return y_pred\n",
    "\n",
    "# test your function\n",
    "list_of_tweets = ['\\033[92m'+'Thank God coronovirus is over', \n",
    "                  '\\033[91m'+'This is seriously ridiculous. Stop hoarding', \n",
    "                  '\\033[92m'+'Government efforts are great.', \n",
    "                  '\\033[93m'+'Nonsense', #anomaly in classification\n",
    "                  '\\033[91m'+'Panic-buying is pushing up prices.#coronavirus', \n",
    "                  '\\033[92m'+'THANK YOU to make hand sanitizer in distillery amid outbreak']\n",
    "\n",
    "for tweet in list_of_tweets:\n",
    "    print( '%s -> %f' % (tweet, predict_custom_tweet(tweet, freqs)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "da8ea1d9",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.11.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}