Sentiment Analysis on UTHM Issues with Big Data

Abstract

Nowadays, social media platform such as Twitter, WhatsApp, Facebook and it Messenger, as well as Instagram plays a very importance role to the society. Twitter is a micro-blogging platform that is able to provide a remarkable amount of data that can be used in several number of sentiment analysis applications such as predictions, reviews, and elections. Sentiment Analysis is a process of extracting information of issues or specific topic from enormous amount of data and categorizes it into different classes. The main target of this project is to classify Twitter data into sentiments value either positive, neutral or negative on data collected regarding Universiti Tun Hussein Onn Malaysia (UTHM) issues. This sentiment was classified using sentiment classifier, while data is trained on a Naïve Bayes Classifier, on TextBlob Python library. Lastly, results were displayed to the user, through a web application using Jupyter Notebook. This study found out that the percentage for positive, neutral and negative tweets regarding UTHM issues were 74%, 26% and 0% in English tweets, meanwhile 17%, 82% and 1 % of Bahasa Melayu tweets, respectively. Positive and neutral sentiments analysis shows positive perception of the products and services, thus promoting and branding UTHM worldwide.