Skip to main content Link Search Menu Expand Document (external link)

Why GeoDesk?

Since its launch in 2004, a vibrant software ecosystem has emerged around OpenStreetMap. However, geospatial analytics remains a difficult task. Most processing tasks are resource-intensive, requiring high-end workstations and in-depth expertise. The OSM project was meant to democratize geospatial data, but these hurdles discourage many potential adopters.

The crux of the problem is this: That shiny OpenStreetMap file isn’t particularly useful until you load it into a database, where you can dissect it with spatial queries. But importing into a traditional SQL-based DBMS turns that already hefty OSM file into a monstrous hulking beast. You’re gonna need a bigger drive! And you’re going to need tons of patience, because that database import will take many hours even on a beefy machine.

So we went back to the drawing board and reimagined data storage. Instead of using a relational database, GeoDesk stores OSM data in a Geographic Object Library (GOL). GOLs have the following advantages:

  • Compact file size: GOLs are stored as single files, which are typically only 40 percent larger than the dataset in .osm.pbf format. This is a small fraction of the footprint of a traditional database.

  • Lightning-fast queries: The most common spatial queries perform fifty times faster than their SQL equivalents.

  • Designed for OSM: Unlike most spatial databases, GOLs store not only the geometries of features, but also support OSM concepts like relations.

  • Simplified distribution of OSM data: Any GOL can be turned into a compressed tile repository, from which users download only the regions they need. This greatly reduces storage and download costs.

  • Easy to use: The GeoDesk API provides a powerful query language based on familiar MapCSS. Results are returned as Java or Python objects — no need for tedious object-relational mapping.

  • Seamless integration with the Java Topology Suite (JTS) (and Python’s Shapely) for advanced vector operations: buffer, generalize, union, convex and concave hulls, triangulation, Voronoi diagrams, and much more.

  • Modest hardware requirements: GeoDesk performs well on just about any system that can run 64-bit Java or Python.

The GeoDesk database engine and the GOL command-line utility are free & open-source.