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

We build an end-to-end spatial graph learning pipeline using city2graph. We collect urban POI and street network data from OpenStreetMap, with a synthetic fallback for reliability.

We engineer spatial features, construct several proximity graph families, and compare how each represents the same urban environment. We then build heterogeneous and homogeneous graphs, convert them to PyTorch Geometric, and train a GraphSAGE model to predict POI categories from spatial structure.