<![CDATA[Consumer reviews on digital platforms store valuable information about consumer preferences and behavior that has not been widely utilized by Micro, Small, and Medium Enterprises (MSMEs). This study aims to identify patterns of consumer preferences for culinary MSMEs in East Surabaya using review data from Google Maps. A total of 306,528 reviews from 106 culinary MSMEs across 7 sub-districts were collected through web scraping techniques, followed by data cleaning, data transformation, and text preprocessing using the NLTK and Sastrawi libraries through five stages: case folding, special character removal, tokenization, stopword removal, and stemming. Consumer satisfaction classification was based on numerical rating values without a full NLP approach, making it more robust to the linguistic variation of mixed Indonesian-language text. The results show an overall average rating of 4.50 out of 5, with 82.3% of sampled reviews classified as positive; keyword-frequency analysis of 350 preprocessed reviews found that food taste and price ("enak", "rasa", "harga") were the most frequently mentioned attributes, while Gubeng sub-district recorded the highest review volume despite having fewer businesses than Rungkut. The contribution of this study lies in combining a quantitative rating-based approach with qualitative keyword-frequency analysis that can be adopted without complex computational infrastructure. This study demonstrates that Google Maps review data can serve as a valid source of information for understanding consumer preference patterns in the culinary MSME segment at the city level.]]>
