The format that shall not be named

Author
Published

February 13, 2025

… for I am not interested in the Shapefile Wars and I do not want to wake up the geospatial format zealots. But: By now, a cultural history of the Shapefile, the world’s foremost geospatial data format1 since the early 1990s (!), would be warranted, such is its impact. James Fee’s article “Shapefile: The Persistent Legacy of a GIS Workhorse” makes the point:

I’m pretty sure Esri never intended the Shapefile to be a long-term standard, its simplicity, compatibility, and widespread adoption have cemented its place in geospatial workflows across industries. (…) Even as more efficient, flexible, and robust alternatives like GeoJSON, Parquet, and GeoPackage have emerged, the Shapefile persists, often by necessity rather than preference.

Luckily, I don’t usually have to cater to workflows that rely on Shapefiles at all2.

As to alternatives, there are many to chose from. I find it hard to name “the best” data format, since it very much depends:

I do find GeoParquet and FlatGeoBuf particulary interesting vector data formats, for some of my use cases – ymmv3, obviously.

Footnotes

  1. much in the same way Excel is the world’s foremost “database” – besides SQLite↩︎

  2. One exception comes to mind: to my dismay, a popular traffic modelling software only supports Shapefile for GIS data exchange.↩︎

  3. your mileage may vary↩︎