Analisis Persepsi Publik Terhadap Provider Telkomsel melalui Tweet dan Komentar di Media Sosial dengan Pemodelan Topik dan Analisis Sentimen Berbasis Emosi

Abstrak

This study aimed to analyze public perception towards Telkomsel on social media using Latent Dirichlet Allocation (LDA) topic modelling and emotion-based sentiment analysis with the NRC lexicon. Data were collected from Twitter tweets and comments on Instagram and Facebook from October 25-31, 2023. The results showed that public perception is dominated by complaints and dissatisfaction related to difficulties in activating data packages, problems with topping up credit, poor network quality, expensive package prices, bugs in the MyTelkomsel application, and slow customer service response. Sentiment analysis revealed 53.8% negative sentiment on Twitter and Instagram, and 66.2% on Facebook. The dominant emotion was anger, with 42.75% on Twitter, 47.88% on Instagram, and 82.22% on Facebook. These findings indicate public dissatisfaction with Telkomsel, particularly regarding network quality, package prices, and customer service. Therefore, it is recommended that Telkomsel improve its service quality by focusing on enhancing network quality, strengthening signals in problematic areas, accelerating the data package activation and credit top-up processes, and improving the responsiveness and effectiveness of customer service.