Twitter, Instagram, Youtube Speak: Understanding Sentiments on LRT Jabodebek Services via Inset Lexicon, IndoBERT and BERTopic Approaches

Abstract

Rapid urbanization in the Jabodetabek region has led to an increased demand for public transportation. Responding to this need, the government has initiated the development of a new public transportation mode, namely the LRT Jabodebek. However, as a new public transportation mode, the LRT Jabodebek has both strengths and weaknesses in serving the community. Various public comments are expressed through social media platforms. To enhance service quality, it is crucial to pay attention to public comments. Therefore, a sentiment analysis is required to identify and delve into both positive and negative sentiments regarding the LRT Jabodebek service through comments on Twitter, Instagram, and Youtube. The methodology involves a combination of Lexicon-based, IndoBERT model, and BERTopic approaches to gain a deeper understanding of the Jabodebek LRT service trends. The study reveals that 55.9% of the 8,523 comments carry a negative sentiment, and the IndoBERT model achieves an accuracy of 85.97% in sentiment classification.