GeoTiles offers ready-made geodata with a focus on the Netherlands.

Nationwide data sets are overwhelming in size, even in a small country like the Netherlands. GeoTiles aims to provide a consistent tiling structure to access and distribute geo-data in a user-friendly way. (Geo)Tiles that allow for both visualisation as well as massive parallel processing. Therefore, large data sets are chunked in smaller tiles and presented to the visitor on a map: lower zoom levels show the main tiles, higher zoom levels reveal the sub-units that require an even finer tiling structure.

GeoTiles began in 2017 as a data access portal for student and research projects, linking to external resources. Since then, derived and enriched data sets were added, and the project grew in size, to approximately 30 TB (September 2020). Since autumn 2020, a map instead of a list provides access to the available products. The website is still aimed at R&D, but welcomes all visitors.

Coloured AHN point clouds

An extensive archive of the point clouds of all four releases of the “Actueel Hoogtebestand Nederland”, optimised for processing and coloured with the publicly available aerial photograph. Links to the accompanying digital elevation models (raster) are provided as well.

Sentinel-2 data cube

New addition is a data cube that provides instant access to Sentinel-2 satellite imagery over the Netherlands.

Other resources

Other resources available include aerial photographs (2016-2020) and tiled InSAR deformation data sets. All data is published in the same tiling schema, to ensure compatibility between data sets.


I would like to thank Roderik Lindenbergh for his support of the project. Furthermore, I would like to thank Nederlands Centrum voor Geodesie en Geo-informatica (NCG) for their support of the Sentinel-2 data cube.

The background maps are provided by PDOK and powered by OpenLayers. Software used in processing: GDAL/OGR, PDAL, make, python (GeoPandas, lxml) and LASTools. Usage analytics are performed using Matomo (formerly Piwik).

Technical considerations

A proxy server with a 1½ TB cache is in place to serve the most frequently requested files fast from a data centre in Germany. However, less popular files have to be fetched from the main archive. The bandwidth from the archive is typically only ⅒ of the bandwidth from the cache server. Dates provided as Last-Modified are accurate on this server.

A limit of 10 parallel connections for each client is enforced by the server. However, concurrent connections are in practice only faster for cached files. Moreover, the service provided is not intended to be included in automated downloads of software packages without prior consultation. If you would like to acquire large portions of the dataset, it might be faster to get in touch and arrange either a direct connection or transfer via hard drive.

Recent changes

  • 2023-01-24: Extended the sheet index, for better coverage of the aerial photograph. AHN4 coverage is now complete.
  • 2022-12-05: Added AHN4 (≃ Groningen, Drenthe, Overijssel, Gelderland)
  • 2022-11-30: Added aerial photograph 2022 (25cm).
  • 2022-07-21: Bestand Bodemgebruik (BBG) 2017 added to the data cube.
  • 2022-02-01: AHN4 coloured subunits of Noord-Brabant and Limburg available.
  • 2021-10-25: reduced performance due to drive failure, performance should be restored within 48 hours.
  • 2021-07-22: tiled versions of AHN1, AHN2 and AHN3 will be updated to include an ‘overlap’ flag on points within the 20 meter buffer with other tiles.
  • 2021-07-06: initial version of the data cube manual.
  • 2021-07-02: first coloured subunits of AHN4 available for download.
  • 2021-06-11: addition of AHN4, including LAX-index and txt-info for all available tiles.
  • 2021-06-03: The severe connection issues that occurred on 2021-05-30 are resolved. The website and archive are operational again.
  • 2021-05-04: tiled InSAR deformation time series from Rijkswaterstaat are now available.
  • 2021-04-19: the aerial photograph of 2020 replaced by a version extracted from the ECW, rather than scraped from WMS.