SENTIMENT ANALYSIS AND CLUSTERING OF ISP SERVICE USERS BASED ON SOCIAL MEDIA PLATFORM X IN INDONESIA USING K-MEANS METHOD

This study analyzes user sentiment towards Internet Service Providers (ISPs) in Indonesia using the social media platform X (formerly Twitter). Data was collected using web scraping and processed with TF-IDF to convert text into numerical representations. Sentiment was determined using the NLP BERT model. The K-Means Clustering method was used to group negative tweets based on content similarity. The data consists of 6.000 negative tweets. The analysis identified three main clusters: technical issues (network disruptions, high prices, slow connections), customer service and interactions, and communication and customer satisfaction. Inter-cluster distances were: Cluster 1 and Cluster 2 (0.460), Cluster 1 and Cluster 3 (0.349), Cluster 2 and Cluster 3 (0.341). Intra-cluster variations were: Cluster 1 (0.140), Cluster 2 (0.113), Cluster 3 (0.064). Managerial implications include the need to improve technical service quality, customer service responsiveness, and billing transparency. The study's limitations include the limited amount of data and potential bias. Future research is suggested to use regression models to predict customer satisfaction based on complaint patterns and user sentiment.