<![CDATA[Palembang's tourism sector increasingly relies on online reviews as visitor satisfaction indicators, yet the large volume of unstructured review data complicates manual analysis. This study aims to analyze visitor satisfaction patterns at Amanzi Waterpark Palembang using K-Means Clustering algorithm on 1,812 Google Maps reviews collected through web scraping techniques. The analytical process includes text preprocessing, TF-IDF weighting, TruncatedSVD dimensionality reduction, and clustering with k=5. Research findings identify five visitor experience segments: Family Recreation (12.4%, rating 4.69), General Positive Reviews (8.9%, rating 4.55), Cleanliness & Comfort (7.1%, rating 4.60), Mixed Reviews & Complaints (67.5%, rating 3.99), and English Language Reviews (4.1%, rating 4.57). Critical findings reveal that 67.5% of reviews fall into the cluster with the lowest rating, dominated by complaints regarding pool water cleanliness, operational system complexity, and perceived high prices. Service quality inconsistency is identified through differing cleanliness sentiments across clusters, indicating standards not consistently maintained especially during peak visit periods . This research provides practical contributions in the form of strategic recommendations for cleanliness improvement, payment system simplification, and quality control consistency, while academically enriching the literature on text mining applications in Indonesia's tourism sector.]]>
