Identifikasi Kawasan Strategis Wisata Kuliner Berbasis Analisis Spasial dan Machine Learning di Kota Semarang

Abstract

Semarang City has significant potential for local culinary tourism, but it lacks an integrated spatial mapping of tourist destinations. This study aims to identify strategic culinary tourism areas through spatial analysis and machine learning based on data from 1,311 restaurants and over 40 indicators of facilities and tourist attractions. The methods used include EDA, the HDBSCAN algorithm optimized by DBCV, and zone assessment based on restaurant density, ratings, facilities, and proximity to tourist attractions. Visualization was performed using the Folium interactive map. The results showed that the majority of restaurants have high ratings (average 4,53), but service digitization remains low. Clustering resulted in three zones: the Gold Zone (842 restaurants), the Silver Zone (34 restaurants), and the Bronze Zone (435 restaurants). Areas such as Banyumanik and Tembalang were identified as potential destinations for thematic development. Recommendations include strengthening the branding of the Gold Zone, promoting the Silver Zone, and digitizing MSMEs in the Bronze Zone. This study demonstrates that spatial approaches and machine learning can support data-driven culinary planning aligned with the Central Java RPJMD.