Real Time Sentiment Analysis of Twitter Data

Abstract

This project does live sentiment analysis of Twitter data to classify it into positive, negative, or neutral sentiments. It focuses on extracting tweets in real-time using Twitter APIs, pre-processing it, and then applying sentiment analysis to it. This data will help us apprehend people's perspectives and opinions about diverse topics. Different machine learning models such as, SVM, Naïve Bayes, Logistic Regression and Random Forest are trained on a relevant dataset to achieve the best possible accuracy. The results of the project are reported for the Random Forest classifier since it resulted in the highest accuracy on testing data. After prediction of the sentiments of real-time tweets the results have been visualized in the form of different types of charts. Nowadays, when everything is on social media, it is crucial to detect and limit negative exposure to keep the social media sites safe to use and positive for all age groups. Twitter is one of the most influential platforms today where a lot of data is generated daily making it the most appropriate platform for our project.