GeoPandas 1.0 and beyond — Martin Fleischmann

Python

Learn about GeoPandas 1.0's major performance gains, new features like GeoArrow support, and future plans including spherical geometry and Rust implementation.

Key takeaways

GeoPandas provides subclasses of Pandas objects - GeoPandas geo series and geo data frame - with added spatial functionality
Version 1.0 brings significant performance improvements:
- Requires Shapley 2.0 with vectorized interface
- Switched from Fiona to Pyagria for faster file reading/writing
- New GeoArrow and GeoParquet support for efficient geometry encoding
- Coverage-based union operations for improved performance
Key dependencies simplified:
- Core: Pandas, Shapley (geometries), Fiona/Pyagria (file I/O), Pyproj (coordinate systems)
- Some dependencies like GDAL made optional to improve installation experience
Major new features:
- D-within spatial predicate for distance-based joins
- Full Shapley API parity
- Interactive visualization support
- Coverage-based operations for topology-aware processing
Future developments planned:
- Spherical geometry support through Spherley (Google S2)
- Curved geometry types support
- Coverage simplification improvements
- Potential Rust implementation for better performance
- Enhanced GeoArrow integration
Project evolution:
- Started in 2013 at SciPy conference
- Received NumFOCUS affiliation in 2020
- Over 200 code contributors to date
- Removed controversial datasets module in favor of neutral geodatasets

GeoPandas 1.0 and beyond — Martin Fleischmann

More talks