Perpustakaan Digital ITB

This study explores the use of Point of Interest (POI) data as a planning support tool within the Detailed Spatial Plan (DSP) framework. While POI data has been widely used in urban research, its integration into regulatory zoning remains limited. Using a case study in Pekanbaru, Indonesia, POI-based zones were developed from Google Maps data and assessed for reliability, spatial distribution, and planning relevance. Functional classification was conducted using the Category-to-Ratio (CR) Index and Simpson Index. To address data sparsity and capture complex spatial patterns, Graph Attention Networks (GAT) were applied to generate synthetic POI embeddings and simulate activity scenarios across different urban contexts. Validation included manual POI checks and assessment of zoning outcomes through green area comparison, alignment with regional plans, and usability via the Analytic Hierarchy Process (AHP), focusing on efficiency, legality, availability, and reliability. Findings demonstrate the value of POI-based zoning for highlighting urban structure and diversity, especially in early planning stages. The model offers a spatially grounded, transferable approach aligned with Planning Support Science, bridging data-driven tools with institutional processes.