<![CDATA[In the digital era, online reviews have increasingly become a valuable source of user-generated language data for understanding users’ experiences, evaluations, and opinions. This study examines public perceptions of Museum X after its post-fire reopening through Google Maps online reviews as the data source. The corpus comprised 1,822 Indonesian-language reviews published from 15 October 2024–1 June 2026. The study combines lexicon-based sentiment analysis using the Indonesian Sentiment Lexicon (InSet) and BERTopic topic modelling to identify lexical polarity and dominant museum visitor experience issues. The results show that positive lexical sentiment was the largest category (46.54%), followed by negative (32.38%), no coverage (17.29%), and neutral (3.79%). The mean rating was high at 4.69 out of 5. A manual validation on a stratified sample (n = 249) yields 81.5% agreement (κ = 0.72) and confirms that no-coverage reviews are predominantly sentiment-bearing. BERTopic produced five regular topics and an outlier cluster, with the dominant topic concerning ticketing, entry, services, collections, and price (820 reviews), while the other topics captured historical-cultural collections, general positive impressions and comfort, family education, and history learning. Combining lexical sentiment, rating, and topic context shows that Museum X’s educational and historical value that visitors express in their reviews remains prominent, while ticketing and entry management require attention. The study demonstrates a transparent, coverage-aware workflow for service evaluation from unsolicited online feedback.]]>
Copyrights © 2026
