GeoAI for city-wide pedestrian flow mapping and urban regeneration

Understanding city-wide pedestrian movement is critical for urban regeneration, but its comprehension is often hindered by data scarcity. This study introduces the Geospatial Pedestrian Flow Engine (GPFE) as a GeoAI framework designed to map and predict pedestrian activity at a city-wide scale. By integrating diverse and heterogeneous spatial datasets and applying the principle of spatial similarity, using anonymised pedestrian counts from Wi-Fi sensors and video cameras integrated with land use, street network, and point-of-interest layers, GPFE delivers reliable estimations of pedestrian flow across entire urban environments, crucially addressing areas with sparse pedestrian data. The resulting framework produces high-resolution flow maps that reveal dynamic spatial and temporal movement patterns, offering actionable insights for infrastructure investment, mobility planning, and demand-driven urban regeneration. Validated through spatial cross-validation across commercial, residential, and recreational zones in Loughborough, UK, the framework supports evidence-based strategies to foster more sustainable, inclusive, and resilient cities.