Globetrotting with data – GeoArrow and GeoParquet

Geospatial data, or data that can be linked to specific geographic locations, has historically always been on the fringes of the data space. Whether you’re modeling shipping routes, weather maps, real estate, or – more recently – modeling the spread of disease, it’s vital to be able to represent and transfer geospatial data across distributed systems. So, why is it so difficult?

• There are many, many, different ways to represent geospatial data. As a result, the coordinate reference system (CRS) must be kept and propagated with the data itself so that it can be joined with data coming from other sources. For reference, a CRS is just a standardized way of interpreting geographic references so that tools and utilities have a uniform way of understanding things.
• The type systems for custom geospatial data formats tend to include several non-spatial types, but the popular general data storage...