Project_NLP_Twitter_US_Airline_Sentiment.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "sTTeOe27C0j7"
   },
   "source": [
    "**Author: Pablo Ulises Hernandez Garces**\n",
    "\n",
    "**link Github: https://github.com/GreatLearningAIML1/gl-pgp-aiml-uta-intl-apr20-UlisesHdzG**\n",
    "\n",
    "\n",
    "**Data Description**\n",
    "\n",
    "A sentiment analysis job about the problems of each major U.S. airline. Twitter data was scraped from February of 2015 and contributors were asked to first classify positive, negative, and neutral tweets, followed by categorizing negative reasons (such as \"late flight\" or \"rude service\").\n",
    "\n",
    "\n",
    "**Dataset**\n",
    "\n",
    "The project is from a dataset from Kaggle. Link to the Kaggle project site: \n",
    "https://www.kaggle.com/crowdflower/twitter-airline-sentiment\n",
    "\n",
    "The dataset has to be downloaded from the above Kaggle website. The dataset has the following columns:\n",
    "\n",
    "* tweet_id\n",
    "* airline_sentiment\n",
    "* airline_sentiment_confidence\n",
    "* negativereason\n",
    "* negativereason_confidence\n",
    "* airline\n",
    "* airline_sentiment_gold\n",
    "* name\n",
    "* negativereason_gold\n",
    "* retweet_count\n",
    "* text\n",
    "* tweet_coord\n",
    "* tweet_created\n",
    "* tweet_location\n",
    "* user_timezone\n",
    "\n",
    "**Objective**\n",
    "\n",
    "To implement the techniques learnt as a part of the course.\n",
    "\n",
    "**Learning Outcomes**\n",
    "\n",
    "* Basic understanding of text pre-processing.\n",
    "* What to do after text pre-processing:\n",
    "  * Bag of words\n",
    "  * Tf-idf\n",
    "* Build the classification model.\n",
    "* Evaluate the Model.\n",
    "\n",
    "**Steps and tasks**\n",
    "\n",
    "1. **Import the libraries, load dataset, print shape of data, data description.**\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 187
    },
    "id": "KVwnlarmCdN1",
    "outputId": "c7801220-ccbd-4f00-ea92-d4804faf2c2f"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: contractions in /usr/local/lib/python3.6/dist-packages (0.0.25)\n",
      "Requirement already satisfied: textsearch in /usr/local/lib/python3.6/dist-packages (from contractions) (0.0.17)\n",
      "Requirement already satisfied: pyahocorasick in /usr/local/lib/python3.6/dist-packages (from textsearch->contractions) (1.4.0)\n",
      "Requirement already satisfied: Unidecode in /usr/local/lib/python3.6/dist-packages (from textsearch->contractions) (1.1.1)\n",
      "[nltk_data] Downloading package stopwords to /root/nltk_data...\n",
      "[nltk_data]   Package stopwords is already up-to-date!\n",
      "[nltk_data] Downloading package punkt to /root/nltk_data...\n",
      "[nltk_data]   Package punkt is already up-to-date!\n",
      "[nltk_data] Downloading package wordnet to /root/nltk_data...\n",
      "[nltk_data]   Package wordnet is already up-to-date!\n"
     ]
    }
   ],
   "source": [
    "!pip install contractions\n",
    "\n",
    "import re, string, unicodedata                          # Import Regex, string and unicodedata.\n",
    "import contractions                                     # Import contractions library.\n",
    "from bs4 import BeautifulSoup                           # Import BeautifulSoup.\n",
    "\n",
    "import numpy as np                                      # Import numpy.\n",
    "import pandas as pd                                     # Import pandas.\n",
    "import nltk                                             # Import Natural Language Tool-Kit.\n",
    "from google.colab import drive\n",
    "\n",
    "nltk.download('stopwords')                              # Download Stopwords.\n",
    "nltk.download('punkt')\n",
    "nltk.download('wordnet')\n",
    "\n",
    "from nltk.stem import PorterStemmer                     # Stemmer\n",
    "from nltk.corpus import stopwords                       # Import stopwords.\n",
    "from nltk.tokenize import word_tokenize, sent_tokenize  # Import Tokenizer.\n",
    "from nltk.stem.wordnet import WordNetLemmatizer         # Import Lemmatizer.\n",
    "\n",
    "from sklearn.feature_extraction.text import CountVectorizer          #For Bag of words\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer          #For TF-IDF\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score, confusion_matrix\n",
    "\n",
    "\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "id": "ZAaSSsO2Role",
    "outputId": "dfacf9a2-678d-4023-f062-d68dcbffbc51"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"
     ]
    }
   ],
   "source": [
    "# loading data\n",
    "drive.mount('/content/drive')\n",
    "project_path = '/content/drive/My Drive/AIML/Project 8 - NLP/Data/'\n",
    "data = pd.read_csv(project_path + 'Tweets.csv')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "id": "qyVKScLCXOE2",
    "outputId": "c3cce06b-036c-4be3-aea2-38324d33e575"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(14640, 15)"
      ]
     },
     "execution_count": 25,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Data shape\n",
    "data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 391
    },
    "id": "jE9k6kLJX9qK",
    "outputId": "9b456f3c-22fb-4c8f-a194-5ceccef68965"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 14640 entries, 0 to 14639\n",
      "Data columns (total 15 columns):\n",
      " #   Column                        Non-Null Count  Dtype  \n",
      "---  ------                        --------------  -----  \n",
      " 0   tweet_id                      14640 non-null  int64  \n",
      " 1   airline_sentiment             14640 non-null  object \n",
      " 2   airline_sentiment_confidence  14640 non-null  float64\n",
      " 3   negativereason                9178 non-null   object \n",
      " 4   negativereason_confidence     10522 non-null  float64\n",
      " 5   airline                       14640 non-null  object \n",
      " 6   airline_sentiment_gold        40 non-null     object \n",
      " 7   name                          14640 non-null  object \n",
      " 8   negativereason_gold           32 non-null     object \n",
      " 9   retweet_count                 14640 non-null  int64  \n",
      " 10  text                          14640 non-null  object \n",
      " 11  tweet_coord                   1019 non-null   object \n",
      " 12  tweet_created                 14640 non-null  object \n",
      " 13  tweet_location                9907 non-null   object \n",
      " 14  user_timezone                 9820 non-null   object \n",
      "dtypes: float64(2), int64(2), object(11)\n",
      "memory usage: 1.7+ MB\n"
     ]
    }
   ],
   "source": [
    "data.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "s3CivwuvYqP5"
   },
   "source": [
    "We can see that there are some variables with null values in the dataset. However, the variables with missing values are not relevant for the objective of this project."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "sO5yN1jwQsKo"
   },
   "source": [
    "2. **Understand of data-columns:** \n",
    "  * Drop all other columns except “text” and “airline_sentiment”.\n",
    "  * Check the shape of data.\n",
    "  * Print first 5 rows of data.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "id": "T6vSgguOQriq"
   },
   "outputs": [],
   "source": [
    "# Dropping all columns except \"text\" and \"airline_sentiment\"\n",
    "data= data.drop(data.columns[~data.columns.isin(['text','airline_sentiment'])], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 221
    },
    "id": "0dO9pwbBajUk",
    "outputId": "d1c1d63a-e1d6-49cf-cce6-18f678213da4"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Data shape: (14640, 2)\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>airline_sentiment</th>\n",
       "      <th>text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>neutral</td>\n",
       "      <td>@VirginAmerica What @dhepburn said.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>positive</td>\n",
       "      <td>@VirginAmerica plus you've added commercials t...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>neutral</td>\n",
       "      <td>@VirginAmerica I didn't today... Must mean I n...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>negative</td>\n",
       "      <td>@VirginAmerica it's really aggressive to blast...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>negative</td>\n",
       "      <td>@VirginAmerica and it's a really big bad thing...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  airline_sentiment                                               text\n",
       "0           neutral                @VirginAmerica What @dhepburn said.\n",
       "1          positive  @VirginAmerica plus you've added commercials t...\n",
       "2           neutral  @VirginAmerica I didn't today... Must mean I n...\n",
       "3          negative  @VirginAmerica it's really aggressive to blast...\n",
       "4          negative  @VirginAmerica and it's a really big bad thing..."
      ]
     },
     "execution_count": 28,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print('Data shape:',data.shape)\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 170
    },
    "id": "P6XqnuczExT1",
    "outputId": "3b329e88-f147-4324-f0b0-20295c06a401"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 14640 entries, 0 to 14639\n",
      "Data columns (total 2 columns):\n",
      " #   Column             Non-Null Count  Dtype \n",
      "---  ------             --------------  ----- \n",
      " 0   airline_sentiment  14640 non-null  object\n",
      " 1   text               14640 non-null  object\n",
      "dtypes: object(2)\n",
      "memory usage: 228.9+ KB\n"
     ]
    }
   ],
   "source": [
    "data.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "eXuZ2oKYEzSt"
   },
   "source": [
    "We can see that the variables of interest do not have null values."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "PPnIFxUmQuYa"
   },
   "source": [
    "3. **Text pre-processing: Data preparation.**\n",
    "  * Html tag removal.\n",
    "  * Tokenization.\n",
    "  * Remove the numbers.\n",
    "  * Removal of Special Characters and Punctuations.\n",
    "  * Conversion to lowercase.\n",
    "  * Lemmatize or stemming.\n",
    "  * Join the words in the list to convert back to text string in the dataframe. (So that each row\n",
    "contains the data in text format.)\n",
    "  * Print first 5 rows of data after pre-processing.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 204
    },
    "id": "g3ZhTWTCQuDZ",
    "outputId": "3696b0f8-0e69-417d-a871-c2014126eb7c"
   },
   "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>airline_sentiment</th>\n",
       "      <th>text</th>\n",
       "      <th>clean_text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>neutral</td>\n",
       "      <td>@VirginAmerica What @dhepburn said.</td>\n",
       "      <td>@VirginAmerica What @dhepburn said.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>positive</td>\n",
       "      <td>@VirginAmerica plus you've added commercials t...</td>\n",
       "      <td>@VirginAmerica plus you've added commercials t...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>neutral</td>\n",
       "      <td>@VirginAmerica I didn't today... Must mean I n...</td>\n",
       "      <td>@VirginAmerica I didn't today... Must mean I n...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>negative</td>\n",
       "      <td>@VirginAmerica it's really aggressive to blast...</td>\n",
       "      <td>@VirginAmerica it's really aggressive to blast...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>negative</td>\n",
       "      <td>@VirginAmerica and it's a really big bad thing...</td>\n",
       "      <td>@VirginAmerica and it's a really big bad thing...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  airline_sentiment  ...                                         clean_text\n",
       "0           neutral  ...                @VirginAmerica What @dhepburn said.\n",
       "1          positive  ...  @VirginAmerica plus you've added commercials t...\n",
       "2           neutral  ...  @VirginAmerica I didn't today... Must mean I n...\n",
       "3          negative  ...  @VirginAmerica it's really aggressive to blast...\n",
       "4          negative  ...  @VirginAmerica and it's a really big bad thing...\n",
       "\n",
       "[5 rows x 3 columns]"
      ]
     },
     "execution_count": 30,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# html tag removal\n",
    "def strip_html(text):\n",
    "    soup = BeautifulSoup(text, \"html.parser\")\n",
    "    return soup.get_text()\n",
    "\n",
    "data['clean_text'] = data['text'].apply(lambda x: strip_html(x))\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 204
    },
    "id": "cDvrJLztRDCA",
    "outputId": "5111a55b-64ca-42db-b373-c896f577852c"
   },
   "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>airline_sentiment</th>\n",
       "      <th>text</th>\n",
       "      <th>clean_text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>neutral</td>\n",
       "      <td>@VirginAmerica What @dhepburn said.</td>\n",
       "      <td>@VirginAmerica What @dhepburn said.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>positive</td>\n",
       "      <td>@VirginAmerica plus you've added commercials t...</td>\n",
       "      <td>@VirginAmerica plus you have added commercials...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>neutral</td>\n",
       "      <td>@VirginAmerica I didn't today... Must mean I n...</td>\n",
       "      <td>@VirginAmerica I did not today... Must mean I ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>negative</td>\n",
       "      <td>@VirginAmerica it's really aggressive to blast...</td>\n",
       "      <td>@VirginAmerica it is really aggressive to blas...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>negative</td>\n",
       "      <td>@VirginAmerica and it's a really big bad thing...</td>\n",
       "      <td>@VirginAmerica and it is a really big bad thin...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  airline_sentiment  ...                                         clean_text\n",
       "0           neutral  ...                @VirginAmerica What @dhepburn said.\n",
       "1          positive  ...  @VirginAmerica plus you have added commercials...\n",
       "2           neutral  ...  @VirginAmerica I did not today... Must mean I ...\n",
       "3          negative  ...  @VirginAmerica it is really aggressive to blas...\n",
       "4          negative  ...  @VirginAmerica and it is a really big bad thin...\n",
       "\n",
       "[5 rows x 3 columns]"
      ]
     },
     "execution_count": 31,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# contractions removal\n",
    "def replace_contractions(text):\n",
    "    \"\"\"Replace contractions in string of text\"\"\"\n",
    "    return contractions.fix(text)\n",
    "\n",
    "data['clean_text'] = data['clean_text'].apply(lambda x: replace_contractions(x))\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 204
    },
    "id": "BnXwGX_nFb9I",
    "outputId": "dab0bed9-3a4b-4451-cba0-0e52f6fda746"
   },
   "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>airline_sentiment</th>\n",
       "      <th>text</th>\n",
       "      <th>clean_text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>neutral</td>\n",
       "      <td>@VirginAmerica What @dhepburn said.</td>\n",
       "      <td>@VirginAmerica What @dhepburn said.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>positive</td>\n",
       "      <td>@VirginAmerica plus you've added commercials t...</td>\n",
       "      <td>@VirginAmerica plus you have added commercials...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>neutral</td>\n",
       "      <td>@VirginAmerica I didn't today... Must mean I n...</td>\n",
       "      <td>@VirginAmerica I did not today... Must mean I ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>negative</td>\n",
       "      <td>@VirginAmerica it's really aggressive to blast...</td>\n",
       "      <td>@VirginAmerica it is really aggressive to blas...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>negative</td>\n",
       "      <td>@VirginAmerica and it's a really big bad thing...</td>\n",
       "      <td>@VirginAmerica and it is a really big bad thin...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  airline_sentiment  ...                                         clean_text\n",
       "0           neutral  ...                @VirginAmerica What @dhepburn said.\n",
       "1          positive  ...  @VirginAmerica plus you have added commercials...\n",
       "2           neutral  ...  @VirginAmerica I did not today... Must mean I ...\n",
       "3          negative  ...  @VirginAmerica it is really aggressive to blas...\n",
       "4          negative  ...  @VirginAmerica and it is a really big bad thin...\n",
       "\n",
       "[5 rows x 3 columns]"
      ]
     },
     "execution_count": 32,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Removing numbers\n",
    "def remove_numbers(text):\n",
    "  text = re.sub(r'\\d+', '', text)\n",
    "  return text\n",
    "\n",
    "data['clean_text'] = data['clean_text'].apply(lambda x: remove_numbers(x))\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "id": "4Bnz4QEHFdtG"
   },
   "outputs": [],
   "source": [
    "# tokenization\n",
    "data['clean_text'] = data.apply(lambda row: nltk.word_tokenize(row['clean_text']), axis=1) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "id": "u54hMiLnQmPA"
   },
   "outputs": [],
   "source": [
    "stopwords = stopwords.words('english')\n",
    "\n",
    "customlist = ['not', \"couldn't\", 'didn', \"didn't\", 'doesn', \"doesn't\", 'hadn', \"hadn't\", 'hasn',\n",
    "        \"hasn't\", 'haven', \"haven't\", 'isn', \"isn't\", 'ma', 'mightn', \"mightn't\", 'mustn',\n",
    "        \"mustn't\", 'needn', \"needn't\", 'shan', \"shan't\", 'shouldn', \"shouldn't\", 'wasn',\n",
    "        \"wasn't\", 'weren', \"weren't\", 'won', \"won't\", 'wouldn', \"wouldn't\"]\n",
    "\n",
    "# Set custom stop-word's list as not, couldn't etc. words matter in Sentiment, so not removing them from original data.\n",
    "\n",
    "stopwords = list(set(stopwords) - set(customlist))   "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 204
    },
    "id": "GMAt7JzUQqIc",
    "outputId": "14a970ab-b821-49d9-d29b-f1a830d9a433"
   },
   "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>airline_sentiment</th>\n",
       "      <th>text</th>\n",
       "      <th>clean_text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>neutral</td>\n",
       "      <td>@VirginAmerica What @dhepburn said.</td>\n",
       "      <td>virginamerica dhepburn say</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>positive</td>\n",
       "      <td>@VirginAmerica plus you've added commercials t...</td>\n",
       "      <td>virginamerica plus add commercials experience ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>neutral</td>\n",
       "      <td>@VirginAmerica I didn't today... Must mean I n...</td>\n",
       "      <td>virginamerica not today must mean need take an...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>negative</td>\n",
       "      <td>@VirginAmerica it's really aggressive to blast...</td>\n",
       "      <td>virginamerica really aggressive blast obnoxiou...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>negative</td>\n",
       "      <td>@VirginAmerica and it's a really big bad thing...</td>\n",
       "      <td>virginamerica really big bad thing</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  airline_sentiment  ...                                         clean_text\n",
       "0           neutral  ...                         virginamerica dhepburn say\n",
       "1          positive  ...  virginamerica plus add commercials experience ...\n",
       "2           neutral  ...  virginamerica not today must mean need take an...\n",
       "3          negative  ...  virginamerica really aggressive blast obnoxiou...\n",
       "4          negative  ...                 virginamerica really big bad thing\n",
       "\n",
       "[5 rows x 3 columns]"
      ]
     },
     "execution_count": 35,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lemmatizer = WordNetLemmatizer()\n",
    "\n",
    "def remove_non_ascii(words):\n",
    "    \"\"\"Remove non-ASCII characters from list of tokenized words\"\"\"\n",
    "    new_words = []\n",
    "    for word in words:\n",
    "        new_word = unicodedata.normalize('NFKD', word).encode('ascii', 'ignore').decode('utf-8', 'ignore')\n",
    "        new_words.append(new_word)\n",
    "    return new_words\n",
    "\n",
    "def to_lowercase(words):\n",
    "    \"\"\"Convert all characters to lowercase from list of tokenized words\"\"\"\n",
    "    new_words = []\n",
    "    for word in words:\n",
    "        new_word = word.lower()\n",
    "        new_words.append(new_word)\n",
    "    return new_words\n",
    "\n",
    "def remove_punctuation(words):\n",
    "    \"\"\"Remove punctuation from list of tokenized words\"\"\"\n",
    "    new_words = []\n",
    "    for word in words:\n",
    "        new_word = re.sub(r'[^\\w\\s]', '', word)\n",
    "        if new_word != '':\n",
    "            new_words.append(new_word)\n",
    "    return new_words\n",
    "\n",
    "def remove_stopwords(words):\n",
    "    \"\"\"Remove stop words from list of tokenized words\"\"\"\n",
    "    new_words = []\n",
    "    for word in words:\n",
    "        if word not in stopwords:\n",
    "            new_words.append(word)\n",
    "    return new_words\n",
    "\n",
    "def lemmatize_list(words):\n",
    "    new_words = []\n",
    "    for word in words:\n",
    "      new_words.append(lemmatizer.lemmatize(word, pos='v'))\n",
    "    return new_words\n",
    "\n",
    "def normalize(words):\n",
    "    words = remove_non_ascii(words)\n",
    "    words = to_lowercase(words)\n",
    "    words = remove_punctuation(words)\n",
    "    words = remove_stopwords(words)\n",
    "    words = lemmatize_list(words)\n",
    "    return ' '.join(words)\n",
    "\n",
    "data['clean_text'] = data.apply(lambda row: normalize(row['clean_text']), axis=1)\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "3ikBKXI8Qwl8"
   },
   "source": [
    "4. **Vectorization:**\n",
    "  * Use CountVectorizer.\n",
    "  * Use TfidfVectorizer."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "id": "cTQaFzkSQwUC",
    "outputId": "2aa55629-31b9-4535-e9f8-581669792227"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(14640, 1000)"
      ]
     },
     "execution_count": 148,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Vectorization using CountVectorizer\n",
    "vectorizer = CountVectorizer(max_features=1000)                # Keep only 2000 features as number of features will increase the processing time.\n",
    "data_countVect = vectorizer.fit_transform(data['clean_text'])\n",
    "\n",
    "data_countVect = data_countVect.toarray()                        # Convert the data features to array.\n",
    "data_countVect.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 149,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "id": "YGdxR3wDUOGH",
    "outputId": "b5c11543-8067-493d-df07-ec89558d36b2"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(14640, 1000)"
      ]
     },
     "execution_count": 149,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Vectorization using TfidfVectorizer\n",
    "vectorizer_tfidf = TfidfVectorizer(max_features=1000)               # Keep only 2000 features as number of features will increase the processing time.\n",
    "data_tfidf = vectorizer_tfidf.fit_transform(data['clean_text'])\n",
    "\n",
    "data_tfidf = data_tfidf.toarray()                             # Convert the data features to array.\n",
    "\n",
    "data_tfidf.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "jqPFP8cKQy7C"
   },
   "source": [
    "5. **Fit and evaluate model using both type of vectorization.** \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 150,
   "metadata": {
    "id": "NVOU1I7rIpjZ"
   },
   "outputs": [],
   "source": [
    "data['label'] = data['airline_sentiment'].map(lambda x: ({'neutral':0,'negative':1,'positive':2}).get(x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 151,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 173
    },
    "id": "zypFZdbxJoFA",
    "outputId": "6c3ee710-0c3c-4824-e1c0-3a368cf57b10"
   },
   "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>col_0</th>\n",
       "      <th>count</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>label</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>3099</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>9178</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2363</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "col_0  count\n",
       "label       \n",
       "0       3099\n",
       "1       9178\n",
       "2       2363"
      ]
     },
     "execution_count": 151,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.crosstab(index=data['label'],columns=\"count\")     "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "TCKOdzAWKEaE"
   },
   "source": [
    "We have an imbalanced dataset with lots of observations in negative sentiment. This will impact in the accuracy of the model. I will use 'class_weight' parameter from the model equal to 'balanced' in order to handle this problem. \n",
    "\n",
    "**Note:** I ran the model without using this parameter and I got the following average scores in cross validation: 73.0% for CountVect and 74.53% for TF-IDF"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 152,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 255
    },
    "id": "SHf83R23LX6s",
    "outputId": "266cb581-ba3b-4ca2-b7c6-2506954d88c6"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(14640, 1000)\n",
      "(14640,)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0        0\n",
       "1        2\n",
       "2        0\n",
       "3        1\n",
       "4        1\n",
       "        ..\n",
       "14635    2\n",
       "14636    1\n",
       "14637    0\n",
       "14638    1\n",
       "14639    0\n",
       "Name: label, Length: 14640, dtype: int64"
      ]
     },
     "execution_count": 152,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "labels= data['label']\n",
    "print(data_countVect.shape)\n",
    "print(labels.shape)\n",
    "data_countVect\n",
    "labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 153,
   "metadata": {
    "id": "URW4TAU6QyOi"
   },
   "outputs": [],
   "source": [
    "# Splitting data into train and test sets\n",
    "# for Countvect\n",
    "X_train_countvect, X_test_countvect, y_train_countvect, y_test_countvect = train_test_split(data_countVect, labels, test_size=0.3, random_state=42)\n",
    "\n",
    "# for TF-IDF\n",
    "X_train_tfidf, X_test_tfidf, y_train_tfidf, y_test_tfidf = train_test_split(data_tfidf, labels, test_size=0.3, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 161,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 312
    },
    "id": "ONucIuVycJ7_",
    "outputId": "960be942-7a1c-403b-837b-5779efc72598"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Average score (%): 75.42\n"
     ]
    }
   ],
   "source": [
    "###########\n",
    "# Fitting model with CountVectorizer\n",
    "###########\n",
    "# Using Random Forest to build model for the classification of reviews.\n",
    "# Also calculating the cross validation score.\n",
    "forest_countVect = RandomForestClassifier(n_estimators=50, class_weight='balanced')\n",
    "\n",
    "forest_countVect = forest_countVect.fit(X_train_countvect, y_train_countvect)\n",
    "\n",
    "#print(forest_countVect)\n",
    "\n",
    "cv_res = cross_val_score(forest_countVect, X_train_countvect, y_train_countvect, cv=20)\n",
    "\n",
    "cv_plot = pd.DataFrame({\"Iteration\":range(1,21),\"Score CV\":cv_res })\n",
    "#plt.plot(np.array(history.history['accuracy']) * 100)\n",
    "plt.plot(cv_plot['Score CV']*100)\n",
    "plt.ylabel('Score (%)')\n",
    "plt.xlabel('Iteration')\n",
    "plt.legend(['train'])\n",
    "plt.title('Score over CV iterations - using CountVectorizer')\n",
    "plt.show()\n",
    "print('Average score (%):',round( np.mean(cv_res)*100,2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 162,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 312
    },
    "id": "08i7l9aO_-nX",
    "outputId": "8599945f-24a6-4452-a713-566f85c445ca"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Average score (%): 75.92\n"
     ]
    }
   ],
   "source": [
    "###########\n",
    "# Fitting model with TF-IDF\n",
    "###########\n",
    "# Using Random Forest to build model for the classification of reviews.\n",
    "# Also calculating the cross validation score.\n",
    "forest_tfidf = RandomForestClassifier(n_estimators=50, n_jobs=4,class_weight='balanced')\n",
    "\n",
    "forest_tfidf = forest_tfidf.fit(X_train_tfidf, y_train_tfidf)\n",
    "\n",
    "cv_res_tfidf = cross_val_score(forest_tfidf, X_train_tfidf, y_train_tfidf, cv=20)\n",
    "cv_plot_tfidf = pd.DataFrame({\"Iteration\":range(1,21),\"Score CV\":cv_res_tfidf })\n",
    "#plt.plot(np.array(history.history['accuracy']) * 100)\n",
    "plt.plot(cv_plot_tfidf['Score CV']*100)\n",
    "plt.ylabel('Score (%)')\n",
    "plt.xlabel('Iteration')\n",
    "plt.legend(['train'])\n",
    "plt.title('Score over CV iterations - using TF-IDF')\n",
    "plt.show()\n",
    "print('Average score (%):',round( np.mean(cv_res_tfidf)*100,2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 163,
   "metadata": {
    "id": "bXFwTl1bGpcJ"
   },
   "outputs": [],
   "source": [
    "# predictions\n",
    "\n",
    "# CountVectorizer\n",
    "y_pred_countvect = forest_countVect.predict(X_test_countvect)\n",
    "\n",
    "# TF-IDF\n",
    "y_pred_tfidf = forest_tfidf.predict(X_test_tfidf)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 164,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 445
    },
    "id": "Bgmsz4FRGjFT",
    "outputId": "6a5b8e8e-1bd1-47ef-de93-1243cee1e039"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f4fa9010860>"
      ]
     },
     "execution_count": 164,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Confusion matrix\n",
    "conf_mat_countvect = confusion_matrix(y_test_countvect, y_pred_countvect)\n",
    "conf_mat_tfidf = confusion_matrix(y_test_tfidf, y_pred_tfidf)\n",
    "\n",
    "df_cm_countvect = pd.DataFrame(conf_mat_countvect)\n",
    "plt.figure(figsize = (10,7))\n",
    "sns.heatmap(df_cm_countvect, annot=True, fmt='g')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 165,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 445
    },
    "id": "W54O6UiA7hWh",
    "outputId": "edd902f9-79e7-495f-ca27-9088d7458b47"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f4fa3078630>"
      ]
     },
     "execution_count": 165,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_cm_tfidf = pd.DataFrame(conf_mat_tfidf)\n",
    "plt.figure(figsize = (10,7))\n",
    "sns.heatmap(df_cm_tfidf, annot=True, fmt='g')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 166,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 85
    },
    "id": "JYyk7pDiAvrM",
    "outputId": "faeb35cb-d527-4422-dcca-cbbc585c54f1"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy countVect model: 0.7711748633879781\n",
      "Recall_score countVect model: [0.5260181  0.88415068 0.62536023]\n",
      "Precision_score countVect model: [0.57055215 0.83742848 0.71617162]\n",
      "F1-score countVect model: [0.54738081 0.86015557 0.66769231]\n"
     ]
    }
   ],
   "source": [
    "# Scores per category: Using CountVect\n",
    "print('Accuracy countVect model: '+ str(accuracy_score(y_test_countvect  ,y_pred_countvect)))\n",
    "print('Recall_score countVect model: ' + str(recall_score(y_test_countvect,y_pred_countvect, average=None)))\n",
    "print('Precision_score countVect model: ' + str(precision_score(y_test_countvect, y_pred_countvect,average=None)))\n",
    "print('F1-score countVect model: ' + str(f1_score(y_test_countvect,y_pred_countvect,average=None)))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 167,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 85
    },
    "id": "Es7x9l4qPqCE",
    "outputId": "b7bc4612-cab7-407a-dd4f-0edfbaafeefe"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy TF-IDF model: 0.7714025500910747\n",
      "Recall_score TF-IDF model: [0.49434389 0.90298507 0.5907781 ]\n",
      "Precision_score TF-IDF model: [0.58894879 0.82100162 0.73873874]\n",
      "F1-score TF-IDF model: [0.53751538 0.860044   0.65652522]\n"
     ]
    }
   ],
   "source": [
    "# Scores per category: Using TF-IDF\n",
    "print('Accuracy TF-IDF model: '+ str(accuracy_score(y_test_tfidf  ,y_pred_tfidf)))\n",
    "print('Recall_score TF-IDF model: ' + str(recall_score(y_test_tfidf,y_pred_tfidf, average=None)))\n",
    "print('Precision_score TF-IDF model: ' + str(precision_score(y_test_tfidf, y_pred_tfidf,average=None)))\n",
    "print('F1-score TF-IDF model: ' + str(f1_score(y_test_tfidf,y_pred_tfidf,average=None)))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "_td99MN-Q1ZN"
   },
   "source": [
    "6. **Summarize your understanding of the application of Various Pre-processing and Vectorization and performance of your model on this dataset.**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "53OKIsi45_0t"
   },
   "source": [
    "In the pre-processing steps I followed, there are some elements that help to improve the performance of the models. For example, removing some stop words related to negative meaning (such as not, couldn't, didn't, etc.) from the stop words set helps the model to identify pretty well the tweets with a negative sentiment. Additionally, this is really important because there are a lot more negative than neutral or positive tweets. Moreover, for the Airline company it is valuable to understand the issues described in the negative tweets in order to solve them. In other words, the problem to solve is understand what the airline is doing bad and how they can improve their service to reduce bad comments and continuing with the good things. If I had not removed those stop words, the model could not have learnt about the negative sentiment of lots of tweets because of the lack of context given by the stop words removed.\n",
    "\n",
    "The rest of the steps of pre-processing are necessary for having a clean dataset, such as html tags removal, removing contractions, punctuation, numbers and non-ascii characters and converting each word to lowercase. Having implemented these steps we have a more interpretable data, in terms of standard English. Additionally, tokenization is needed to implement all those steps mentioned before to each word.\n",
    "\n",
    "The last part of the pre-processing I implemented was lemmatization. This is a key part of the process. Due to this, models could perform well. In contrast, if I had implemented stemming, performance of models would have been affected because some context had been lost (a disadvantage of stemming compared to lemmatization). Additionally, running time has not been affected for doing lemmatization. \n",
    "\n",
    "On the other hand, vectorization plays an important role in natural language processing. It is a step that we must do in order to feed our models with data and that computers could interpret it. In my process, I implemented both CounVect and TF-IDF vectorizations and also play with some different numbers of max features, I found that with too low or too high numbers of features perform bad in my models, so I kept only 1000 features that seems to be a good number for this problem. And this can be explained as the model could not learned from very few features because the complexity of the problem requires more features(distinct words) but neither from a very large number of distinct words because we would have so many features than the model could not differentiate one \n",
    "observation from another (we would have a sparse matrix).\n",
    "\n",
    "In terms of the performance of the models, both of them (countVect and TF-IDF) had good scores. However, countVect performs a little more better than TF-IDF in almost all the metrics I showed, but TF-IDF predicted correctly more negative sentiment tweets thant countVect."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 195,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 198
    },
    "id": "Su_9WOjgz83v",
    "outputId": "2cbb13ef-1ec1-4aeb-a2fa-399c6bc18d35"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from wordcloud import WordCloud, STOPWORDS, ImageColorGenerator\n",
    "\n",
    "# filtering those tweets with negative sentiment of the test set in order to visualize what are the problems\n",
    "index = y_pred_tfidf==1\n",
    "negative_tweets= X_test_countvect[index]\n",
    "aux = vectorizer_tfidf.inverse_transform(negative_tweets)\n",
    "\n",
    "for i in range(0,negative_tweets.shape[0]):\n",
    "  aux[i]=' '.join(aux[i])\n",
    "words = \" \".join(aux)\n",
    "# Create and generate a word cloud image:\n",
    "wordcloud = WordCloud().generate(words)\n",
    "\n",
    "# Display the generated image:\n",
    "plt.imshow(wordcloud, interpolation='bilinear')\n",
    "plt.axis(\"off\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "TyfeGjYf-4NJ"
   },
   "source": [
    "Finally, from the above wordcloud we can see what are the most frequent words that tweets with negative sentiment have. We can see some related to time like \"hour\", \"delay\", \"late\", etc. Additionally, there are some others related to the service like \"help\", \"book\", \"change\", etc. So, these words could help airline to understand that most of the complains of the customers are related to the time and the quality of the service in the desk or in the call center. With this, airline could implement some strategies to improve their customer services and strictly follow  the schedules."
   ]
  }
 ],
 "metadata": {
  "colab": {
   "collapsed_sections": [],
   "name": "Project NLP - Twitter US Airline Sentiment.ipynb",
   "provenance": []
  },
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}