Traditional urban planning methods for allocation of social urban services–such as schools, hospitals or clinics–rely on static master plans and often fail to adapt to dynamic demographic shifts and localized demand. Our paper presents a data-driven framework to automate location selection for such services, integrating geospatial modeling, open-source datasets, and algorithmic optimization. Leveraging OpenStreetMap and Python libraries (GeoPandas, OSMnx), the tool decomposes cities into interconnected urban blocks, evaluates demand-supply gaps, and prioritizes sites based on regulatory compliance, accessibility, and spatial feasibility. A case study in St. Petersburg’s Vasileostrovsky District demonstrated the tool’s effectiveness: identifying a critical shortage of school seats, it recommended a 1,500-seat school in an underserved block, improving the district’s service coverage coefficient by 16.3%. This outcome outperformed the city’s master plan proposal by 9.5%, highlighting the limitations of conventional top-down approaches. The framework’s block-level granularity enables hyperlocal analysis, ensuring equitable resource allocation while minimizing costs. The study underscores the potential of computational tools to modernize urban governance, offering scalable solutions for cities grappling with population growth and infrastructure demands. Future work plans include expanding the tool to emergency and recreational services, and integrating real-time IoT data.