Implementation of Web Scraping and Data Mining for Performance Evaluation of PT Ceria Multimedia Services

This research aims to implement web scraping techniques to collect testimonial data from the CeriaMultimedia website and perform sentiment analysis to evaluate service quality. The collected data consists of a limited number of testimonials, which are then processed through text preprocessing stages including case folding, tokenizing, filtering, and stemming. The sentiment classification process is conducted using machine learning methods based on TF-IDF weighting and classification algorithms. Due to the limited dataset, the analysis results are used primarily to demonstrate the implementation process rather than to draw generalized conclusions. The results show that the sentiment categories obtained include positive, negative, and neutral sentiments, although not all categories consistently appear in the testing phase. This research highlights the effectiveness of web scraping and text processing techniques while also indicating the need for a larger dataset to improve evaluation accuracy in future studies.