Evaluating Telecom Providers in Malaysia Using Sentiment Analysis and Topic Clustering

Abstract

Malaysian telecommunications industry receives massive amounts of unstructured customer responses on digital platforms but has no systematic processes to derive actionable insights out of such data. The paper introduces a hybrid sentiment classification and topic clustering model to assess four large Malaysian telecom companies: Maxis, CelcomDigi, Digi, and U Mobile based on multi-source user-generated content (UGC) on Reddit and Google Play Store. The suggested pipeline combines the TF-IDF vectorization with the K-Means clustering approach to unsupervised topic discovery with the VADER (Valence Aware Dictionary and Sentiment Reasoner) lexicon-based model of polarity classification of 114,582 customer reviews between 2020 and 2025. A comparative sentiment distribution, temporal trends, regional sentiment, and dominant themes of complaints are visualized in an interactive Power BI dashboard- MyTelcoSentiment-across the seven identified service clusters. The usability assessment using the System Usability Scale (SUS) gave an average of 77.5 which is an indicator of good usability. Findings reveal that the most discussed service dimensions are App and Technical Issues, Internet Speed and Performance, and Overall Satisfaction. Celcom has the highest number of reviews whereas Digi had the greatest positive sentiment ratio. The framework presents a scalable, data-based instrument to telecom providers, consumers and the regulatory authorities like the Malaysian Communications and Multimedia Commission (MCMC) to track and enhance the quality of services.