[P] A Python library processing geospatial data for GNNs with PyTorch Geometric

I’d like to introduce City2Graph, a Python library that converts geospatial data into tensors for GNNs in PyTorch Geometric.

This library can construct heterogeneous graphs from multiple data domains, such as

• Morphology: Relations between streets, buildings, and parcels
• Transportation: Transit systems between stations from GTFS
• Mobility: Origin-Destination matrix of mobility flow by people, bikes, etc.
• Proximity: Spatial proximity between objects

It can be installed by

pip install city2graph

conda install city2graph -c conda-forge

For more details,

• 💻 GitHub: https://github.com/c2g-dev/city2graph
• 📚 Documentation: https://city2graph.net