12+ Geospatial Extensions for PostgreSQL

PostGIS is the de-facto spatial extension for PostgreSQL — adds geometry and geography column types, spatial indexes (GiST, SP-GiST, BRIN), and ~300 ST_* functions implementing the OGC Simple Features standard. It's BSD-licensed, ~2k GitHub stars, used by NASA, OpenStreetMap, and most mapping platforms. PostGIS is mature (first released 2001) and supported by every major managed Postgres provider including AWS RDS, Aurora, Supabase, Neon, and Azure Database for PostgreSQL.

Plain lat/long columns work for storing locations and computing simple distances with the haversine formula. PostGIS becomes essential when you need: spatial indexes (queries like 'find all points within 5 km of X' run in milliseconds instead of full scans), polygon operations (geofencing, area calculations), projections (translating between coordinate systems), or anything beyond point-to-point math. The performance difference at scale is dramatic — a spatial GiST index turns a 10-second sequential scan into a 5ms indexed lookup.

On managed services: PostGIS is pre-installed on RDS, Aurora, Supabase, Neon, and Azure — just run CREATE EXTENSION postgis; as a user with CREATE privilege. On self-hosted Postgres: install the OS package (e.g. apt install postgresql-17-postgis-3 or yum install postgis34_17), then run CREATE EXTENSION postgis; in the target database. Optional companion extensions: postgis_topology, postgis_raster, postgis_tiger_geocoder for US address geocoding.

Every major managed provider supports PostGIS natively, no special config required. AWS RDS supports PostGIS on all current PostgreSQL versions (with raster support on RDS 13+). Aurora supports it identically. Supabase ships PostGIS enabled by default on every project. Neon includes PostGIS in their default extension list. Companion extensions like pgrouting and H3 have narrower support — check the provider's current extension allow-list for those.

geometry uses planar (Cartesian) coordinates — fast computation but inaccurate over long distances on Earth's curved surface. geography uses spherical coordinates and computes great-circle distances on the WGS84 ellipsoid — slightly slower but accurate for global data. Rule of thumb: use geography when your data spans countries or you care about real-world distances (e.g. delivery radius in km). Use geometry when data is in a single projected coordinate system (city, state, regional) and you've already projected the coordinates appropriately.