PySAL was designed with the focus of performing vector-based spatial analysis and therefore it didn't have tools to handle input-output of large raster data. In recent years several geographic data organizations started releasing data in raster format which earlier came in vector format mostly because of advancement in computational capabilities and high storage availability. This led to an increase in the demand for the functionality offered by PySAL to make it work with raster data.

Taking this into consideration, my main aim is to design and implement a lightweight interface which will provide the functionality for streamlining raster data access and making it more accessible to build the data structure accepted by the analytical methods of the PySAL library (mainly libpysal.weights.W/WSP objects) from accessed raster data (which will be an instance of xarray.DataArray). Ultimately, this functionality will open up the use of analytical methods like esda, spatial regression over raster data.



Mragank Shekhar


  • Stefanie Lumnitz
  • Dani Arribas-Bel
  • Levi John Wolf