AI-ASSISTED GEOGRAPHIC ASSESSMENT OF BUSINESS LOCATIONS

Abstract

This article presents an artificial intelligence-assisted geographic decision-support prototype for evaluating potential business locations in Azerbaijan. The study combines a local GeoJSON dataset of 11,807 objects and 69 object types, district-level demographic indicators for Baku, and a Python-based spatial scoring model implemented in a Streamlit application. A large language model is used only to transform a natural-language business idea into structured intent, including target business categories and supporting facility types. The final location recommendation is not generated by the language model, it is produced through deterministic calculations based on Haversine distance, competitor density, supporting-object availability, local demand proxy and demographic advantage. The prototype groups nearby coordinate into grid-based candidate areas, ranks them by a transparent multi-factor score, and presents the results with map visualization and explanatory evidence. The study demonstrates that semantic AI can strengthen business-location analysis when it is combined with verifiable geographic data and explainable mathematical logic.