Whether looking at trends across location or augmenting your data sets with the copious available data, the ability to handle spatial data is a key skill for a data scientist. Unfortunately, accessing this data has a steep learning curve. This talk aims to give a background and basic tooling to get started working with spatial data in python. Examples will be drawn from agriculture in East Africa with an emphasis on leveraging publicly available data.

Background theory

Though it is often possible to have success without delving deep into the theory behind coordinate reference systems, having a familiarity with the concepts makes understanding the tools much easier. We'll define the terms and data structures seen in the often opaque user docs as well as a very brief diversion into mapping and projection theory.

Python tools

Discuss briefly the tools available in python focusing on

• Fiona - reading and writing spatial files
• Shapely - manipulation of geometric objects such as calculating areas, distances, etc.
• Rasterio - reads and writes geospatial raster images
• Geopandas - a high level library that provides a Pandas like interface for spatial data
• Cartopy - plotting for geospatial data