Geopandas - Dask bridge to scale geospatial analysis
- Mentors
- Stefanie Lumnitz, Martin Fleischmann, Brendan Ward, Joris Van den Bossche
- Organization
- NumFOCUS
Geopandas is one of the most popular Python projects for handling geographic problems and data. Whilst Geopandas is useful for handling small data problems, it does not scale well because its geometric computation is single-threaded. Therefore, advanced geospatial analysis is still dependent on the use of relational databases and spatial extensions such as postgreSQL/PostGIS. This project will extend Geopandas capabilities to handle big problems by using Dask to bridge this gap, allowing geometric computations to leverage all available CPU cores. This will be achieved by contributing to the dask-geopandas library, to enable distributing spatial indexing, related spatial partitioning and parallised IO. Additionally, we plan to integrate other spatial operations from Geopandas and basic plotting features.