A Coding Implementation on Spatial Graph Neural Networks for Urban Function Inference Using city2graph, OSMnx, and PyTorch Geometric

An end-to-end spatial graph learning pipeline has been developed using city2graph, OSMnx, and PyTorch Geometric to infer urban functions. The implementation involves collecting urban point of interest (POI) and street network data, engineering spatial features, and constructing various proximity graph families. The approach utilizes a GraphSAGE model to predict POI categories based on the spatial structure, providing insights into effective graph representations for urban analysis.