Community, Sentiment, and Topic Analysis of Public Discourse on the Issue of DPR Allowance Increases on Platform X

Abstract

This study examines public discourse surrounding the issue of the Indonesian House of Representatives (DPR) allowance increase on Platform X (formerly Twitter) by integrating Social Network Analysis (SNA), sentiment analysis using IndoBERT, and topic modeling with Latent Dirichlet Allocation (LDA). Tweets posted between 20 and 27 August 2025 were collected, preprocessed using standard text-mining techniques, and mapped into digital communities through the Louvain community detection algorithm. The analysis identifies ten major communities with distinct thematic orientations, ranging from personalized criticism of DPR leadership and documentation of protest activities to concerns over accountability and more radical narratives, including calls for the dissolution of the DPR. Sentiment analysis reveals that discourse across all communities is predominantly negative, albeit with varying degrees of intensity. Furthermore, topic modeling demonstrates that the allowance increase functions as a catalyst for broader public dissatisfaction related to political representation, institutional transparency, and legislative performance. Overall, this study offers a comprehensive account of how public opinion is structured and disseminated within online social networks, providing empirical insights that may inform policymakers and media organizations in addressing politically sensitive issues.