Sentiment Analysis of Twitter Data on the 2024 Indonesian Presidential Election Using BERT

Abstract

Social media platforms, particularly Twitter, are frequently employed by individuals to articulate their opinions on various subjects in textual form. The proliferation of viewpoints from diverse sources can influence public perceptions on these topics. The greater the popularity of a topic, the more abundant the opinions generated. Currently, the most widely discussed topic is the 2024 Indonesian presidential election. Sentiment analysis, or opinion mining, is an academic discipline that examines sentiments towards a given entity, while text mining involves the extraction of information through processing, classifying, and analyzing extensive datasets. This study will utilize data crawling techniques to gather data from Twitter which will subsequently undergo preprocessing and cleaning. Following this, the cleaned data will be classified by sentiment (positive, negative, or neutral) using a pre-trained language model (BERT) and Natural Language Toolkit (NLTK). The classified data will then be visualized with tools such as Matplotlib and Wordcloud to elucidate the data distribution.