Skip to main content

Accessing the Overture Catalog

Overture publishes a full catalog of aassets in each monthly release: global map data, vector tiles, a STAC catalog, a GERS registry, bridge files, and a changelog. Everything is hosted on Amazon S3 and Microsoft Azure Blob Storage.

Published datasets​

Overture distributes its core map data as GeoParquet files, a column-oriented spatial data format optimized for cloud-native queries. You can scan across the files and pull only the data you need without downloading everything. See this guide from the Cloud Native Geospatial Forum for more on GeoParquet.

ProviderPath
Amazon S3s3://overturemaps-us-west-2/release/<RELEASE>
Microsoft Azurehttps://overturemapswestus2.blob.core.windows.net/release/<RELEASE>

Find the latest <RELEASE> value in the STAC catalog or the STAC viewer. See the quickstart for how to query the catalog programmatically.

Path structure​

The base paths above point to the core map datasets in a release. Overture partitions the data by theme and type; specifying those directories narrows the data you access.

Example S3 path to the infrastructure feature type in the base theme:

s3://overturemaps-us-west-2/release/2026-02-18.0/theme=base/type=infrastructure/*.parquet

Path components:

  • <RELEASE>: date-based version in the format yyyy-mm-dd.x. An archive of releases is maintained on both S3 and Azure.
  • <theme>: one of Overture's six themes — addresses, base, buildings, divisions, places, and transportation.
  • <type>: a feature type within a theme, e.g. infrastructure within base.
  • *.parquet: the * indicates all Parquet files in the directory.

Theme and type mapping​

ThemeFeature types
Addressesaddress
Basebathymetry, infrastructure, land, land_cover, land_use, water
Buildingsbuilding, building_part
Divisionsdivision, division_area, division_boundary
Placesplace
Transportationsegment, connector

Accessing the files​

The Python CLI and DuckDB are the most common ways to query and download Overture data. Both read directly from the cloud and transfer only the data you need. See the quickstart to get started with either tool.

For bulk downloads, use the AWS CLI or AzCopy to copy full Parquet directories to your local machine.

AWS CLI — download all files in the infrastructure type:

aws s3 cp --no-sign-request --recursive \
  s3://overturemaps-us-west-2/release/2026-02-18.0/theme=base/type=infrastructure/ .

AzCopy — download all place data from Azure:

azcopy copy \
  "https://overturemapswestus2.dfs.core.windows.net/release/2026-02-18.0/theme=places/type=place/" \
  "<local directory path>" --recursive

You can also access Overture data through Amazon Athena, Microsoft Synapse, Sedona, and other tools. See the examples for walkthroughs.

PMTiles​

Overture generates PMTiles vector map tiles with each release. These tiles power the Explorer site and are designed for data inspection rather than production cartography.

ProviderPath
Amazon S3s3://overturemaps-extras-us-west-2/tiles/<RELEASE>/<THEME>.pmtiles
HTTPhttps://overturemaps-extras-us-west-2.s3.us-west-2.amazonaws.com/tiles/<RELEASE>/<THEME>.pmtiles

Preview any PMTiles URL at pmtiles.io. To create your own tiles from Overture data, see the PMTiles example and the overture-tiles repository.

STAC catalog​

Overture's STAC catalog is the machine-readable index for all releases. It always points to the latest release and includes metadata for every theme and type: spatial extents, feature counts, column names, links to GeoParquet files on AWS and Azure, and links to PMTiles. Browse it in the STAC viewer.

GERS registry, bridge files, and changelog​

Each release includes artifacts from the Global Entity Reference System (GERS):

  • GERS registry: maps stable GERS IDs to current feature data across releases.
  • Bridge files: tracks how GERS IDs change between releases — additions, removals, and geometry updates.
  • Data changelog: records schema and data changes across releases.

Data mirrors​

Several partners maintain mirrors on these platforms:

These are community-maintained resources that may offer different access patterns or platform-specific tooling.

Last updated on