Twitter Sentiment Analysis

An introduction to Natural Language Processing (NLP) using Analytics Vidhya's Twitter Sentiment Analysis Practice Problem.

Introduction

This project was a learning experience to become familiar with machine learning tools and libraries in Python. The code in this project was not developed by myself, but was instead sourced from this guide.

The purpose of the project was to train a model to predict the sentiment of a tweet. That is, the model was developed to determine whether or not a tweet was racist/sexist.

To complete this task, a Bag-of-words model was created to meaningfully convey the text in a tweet numerically so that a logistic regression model could be used on the data. The final model was able to predict the sentiment of a tweet with an f1 score of 0.53.

Running The Program

This project was created using Jupyter Notebook and Anaconda. The following libraries were utilized:

import re
import pandas as pd
import numpy as np
import string
import nltk
import warning
import matplotlib.pyplot as plot
import seaborn

Acknowledgments

• Analytics Vidhya Twitter Sentiment Analysis Challenge
• Analytics Vidhya Comprehensive Hands on Guide to Twitter Sentiment Analysis