Public Opinion on UK Public Transportation Through Sentiment Analysis and Topic Modeling
Aslıgül Aksan, Hatice Camgöz Akdağ · 2023
Social media has become a valuable data source for gathering and analyzing public opinion on products and services. Among the popular social media platforms, Twitter stands out for its ability to provide place-time information in a text format called tweets. In this study, sentiment analysis and topic modeling of tweets related to public transportation in the United Kingdom were analyzed. Using the Robustly Optimized BERT Pretraining Approach (RoBERTa), tweets are divided according to their polarities: positive, neutral, and negative. Additionally, Latent Dirichlet Allocation (LDA) is applied to positive and negative tweets, and topics providing the causes are obtained. These topics reveal the strengths and weaknesses of the United Kingdom's public transportation service.