GIS-Based Multi-Criteria and Machine Learning Assessment of Rooftop Photovoltaic Deployment Potential in Selected Nigerian Cities

Nigeria's abundant solar resource contrasts sharply with persistent electricity shortages, underscoring the need for robust spatial frameworks to guide rooftop photovoltaic (PV) deployment. This study developed an integrated Geographic Information System (GIS)-based approach for assessing rooftop solar suitability across twelve selected Nigerian cities by combining satellite-derived Global Horizontal Irradiance (GHI), Global Human Settlement Layer (GHSL) data, rooftop geometry proxies, and a density-based shading index. A multi-criteria suitability index was constructed using the Analytic Hierarchy Process (AHP) and validated through Random Forest regression to evaluate model reliability. Results reveal a pronounced north–south gradient in rooftop PV suitability, with northern cities exhibiting greater potential owing to higher solar irradiation and lower atmospheric attenuation. GHI (0.504) and rooftop area (0.245) emerged as the most influential suitability criteria, while urban density and shading significantly moderated local photovoltaic potential. The validation model demonstrated high predictive accuracy (R² = 0.87, RMSE = 0.041, MAE = 0.032), confirming the robustness of the proposed framework. Spatial suitability mapping further highlighted substantial intra-urban variability, emphasizing the importance of integrating urban morphology into rooftop solar assessments. The proposed framework provides a scalable geospatial decision-support tool for prioritizing rooftop PV investments and supporting evidence-based renewable energy planning in Nigeria and other data-constrained urban environments.