<![CDATA[Social media has become the main space for the public to voice their opinions and criticisms of government institutions. One issue that has received a lot of attention is the 17+8 demonstration demanding accountability from the Indonesian House of Representatives (DPR RI) as a legislative body. This study aims to map public opinion of the DPR RI during this momentum using a BERTopic-based topic modeling approach. The research data consists of 21,195 comments from YouTube, TikTok, Instagram, and Facebook. The analysis process included scraping, text preprocessing, semantic representation formation using BERT embedding, dimension reduction with UMAP, and clustering using HDBSCAN. The modeling results produced more than 90 topics covering key issues such as corruption, the role of students, legislative performance, and security forces. Model evaluation showed an intra-topic cohesion value of 0.778, inter-topic separation of 0.412, and topic diversity of 0.749, indicating that the topics formed were coherent, diverse, and contextual. This study proves that BERTopic is quite effective in analyzing informal public opinion and can be used as a basis for developing digital political communication studies in Indonesia.]]>
Copyrights © 2026
