Mapping Public Discourse and Sentiment on Biomass Energy Transition in Indonesia

The transition toward renewable energy requires not only technological and economic assessment but also an understanding of how biomass pathways are discussed in digital public spaces. This study maps thematic patterns and sentiment in Indonesian YouTube discourse on the biomass energy transition. The final corpus comprised 12,168 Indonesian-language comments from 232 eligible videos. Latent Dirichlet Allocation identified two substantive topics among 11,438 topic-assigned comments: Biomass Technologies, Materials, and Practical Applications (n = 6,035; 52.76%) and National Governance, Energy Policy, and Public Affairs (n = 5,403; 47.24%); 730 comments were retained as NO_TOPIC. Sentiment was classified using a domain-specific fine-tuned IndoBERT model, which achieved 82.67% accuracy, macro-F1 = 0.7245, and weighted-F1 = 0.8196 on a 300-comment held-out final test set. Across the full corpus, 7.03% of comments were negative, 67.71% neutral, and 25.26% positive. Negative sentiment was more prevalent in governance- and policy-related discourse (12.42%) than in technical/practical discourse (2.98%). Topic and sentiment were significantly associated, χ²(2, N = 11,438) = 382.17, p < 0.001, Cramer's V = 0.183. These findings describe thematic and affective differentiation within the analyzed YouTube corpus.