Model Komunikasi Kebencanaan Berbasis Analisis Sentimen dan Jaringan Sosial di Twitter dan Instagram Untuk Mitigasi Banjir di Kota Makassar

Flood disasters in Makassar City remain a recurring problem that not only disrupts daily activities but also triggers widespread public dissatisfaction. This study examines disaster communication on social media, specifically Twitter and lnstagram, using sentiment analysis and social network analysis (SNA). A dataset of 6,258 posts and comments related to #BanjirMakassar and #MakassarSiagaBanjir was collected and processed through normalization, fine-tuned lndoBERTweet for sentiment classification, and Gephi for SNA. Results indicate that negative sentiment dominates online conversations, particularly on lnstagram, with significant spikes during ma}or floods in 2023 and 2025. The analysis also reveals that discussions are concentrated on key districts such as Manggala, Tallo, Tamalanrea, and Panakkukang, while hashtags like #banjirmakassar serve as central nodes in the network. lnstitutional and community accounts, including polri presisi and local volunteers, act as amplifiers of information flow. lntegrating sentiment and network findings, this study proposes a participatory disaster communication model where citizens function as co-producers of information and government agencies serve as facilitators and validators. Social media thus becomes not only a platform for expression but also a citizen-driven early warning system that strengthens coordination, solidarity, and responsiveness in flood mitigation.