OPINI NETIZEN PADA CITRA AQUA PASCA SIDAK KDM (ANALISIS TOPIK MODELLING, SENTIMEN DAN EMOSI)

The AQUA communication crisis occurred when Kang Dedi Mulyadi (KDM) made a surprise inspection visit to the Aqua factory in Subang, West Java, in October 2025. The discovery that Aqua's water source came from deep wells, not mountain springs as portrayed in the brand's campaign, triggered a public controversy. The debate expanded into environmental and social issues, including concerns about the impact of groundwater exploitation and land subsidence. AQUA responded with a technical clarification and objections to the term “bore well” used by KDM, which was seen as damaging the brand's reputation. KDM refused to retract his statement to maintain his integrity. The conflict then escalated into a framing battle over credibility and trust. This study aims to map the topics, sentiments, and emotions in the public discourse (netizen comments on YouTube) related to the KDM and AQUA controversy, and to explain why the discourse on this crisis rapidly intensified. The study uses computational text analysis on 13,254 YouTube comments from three popular videos from news portal accounts registered with the Indonesian Press Council, covering the period from October 24–November 17, 2025. Data was collected via the YouTube API and processed (deduplication, tokenization, stopwords removal, stemming, etc.). The analysis includes sentiment classification based on Transformer models (IndoBERT, XLM-R, mBERT), emotion detection based on Plutchik’s taxonomy (8 categories), and topic modeling (LDA and NMF with coherence evaluation). The results of this study show that the public discourse is dominated by negative sentiment (46.9%), followed by neutral (36.8%) and positive (16.3%). The primary negative emotions identified are anger, disgust, and fear, reflecting moral outrage, disappointment over alleged manipulation, and concern about threats. However, temporal analysis shows that positive emotions such as joy and trust also consistently stand out, indicating space for expressions of support and humor amid the controversy. Topic modeling revealed three dominant topics: (1) environmental issues (concerns over the impact of groundwater exploitation and ecological damage), (2) consumer trust issues (the discrepancy between advertising claims and company practices). From the perspective of Rhetorical Arena Theory (RAT), this crisis developed as a multivocal rhetorical arena where the public, political figures, and the company competed to define the issue. The dominance of negative sentiment, as well as emotions of anger and disgust, shows intense negotiation of claims and delegitimization in the digital public sphere. From the Social-Mediated Crisis Communication (SMCC) perspective, KDM acted as the crisis trigger and information disseminator, while AQUA’s response sought to defuse the situation through clarifications and scientific references. The combination of RAT and SMCC explains why the discourse quickly intensified. The interaction between influencers, the public, and the nature of social media accelerated the spread of information, strengthened opposing positions, and shaped a collective opinion that could not be easily changed by technical clarifications alone. The Aqua vs KDM controversy on YouTube represents a digital communication crisis characterized by critical public discourse and diverse emotions. The use of computational text analysis successfully mapped these dynamics, providing insights for PR practitioners to develop crisis communication strategies that consider key issues, dominant sentiments, and the spectrum of public emotions.