Sentiment Analysis of User Reactions to Meta's Threads Launch and Twitter's X Renaming

Abstract

The massive dataset generated by social media platforms has attracted researchers interested in sentiment analysis. Natural language processing (NLP) has become popular for retrieving useful knowledge from user-generated content. In this chapter, the authors conducted a sentiment analysis of comments from Meta's new product, Threads, in competition to Twitter. Authors have tested the system with real-time comments. A dataset of 12,000 comments was extracted from YouTube using the YouTube API containing keywords such as “Threads,” “Twitter,” and “X.” The authors used DistilBERT to weigh the comments and segregate them into positive and negative. Machine learning algorithms were used to analyze the weighted data, and logistic regression was found to outperform random forest and KNN in providing an accuracy of 85.95%. The study highlights the importance of sentiment analysis in understanding user-generated content and its potential applications in the digital era.