Data

This data product contains the classification of the entire building stock of 10 European countries, geographically split based on NUTS1 regions (2024).

Available data

Below is an overview of the available data.

Coverage

The current coverage includes Austria, Belgium, Czechia, France, Germany, Lithuania, the Netherlands, Poland, Slovakia, and Spain. Additional European countries are expected in 2026.

Releases

The dataset is currently split in two releases. The first one, from September 2025, covers Austria, Czechia, Germany, Lithuania, Poland, and Slovakia. The second, from November 2025 covers Belgium, France, the Netherlands, and Spain (withut Basque region). The countries from the latter release are assgined to the existing taxonomic tree (as of now).

Contents

Structure of all releases:

  • {NUTS1}.parquet / {NUTS1}.gpkg - Hierarchical Morphotope Classification results linked to building geometry (available as both Parquet and GPKG)
  • {NUTS1}_morphotopes.parquet / {NUTS1}_morphotopes.gpkg - Hierarchical Morphotope Classification results linked to morphotope geometry (available as both Parquet and GPKG)
  • {NUTS1}_data.parquet - Underlying morphometric data linkable to building geometry
  • morphotope_data.parquet - Aggregated data linkable to individual morphotopes

Additional files available in the initial release (2025-09):

  • label_name.json - Names of the cluster labels for the first three levels of taxonomy
  • pen_portraits.json - Descriptions of the clusters for the first three levels of taxonomy
  • ward_linkage.npy - The taxonomy itself encoded as a scipy linkage object

Additionally, the repository managed by Charles University contains PMTiles representation.

  • complete.pmtiles - Hierarchical Morphotope Classification results linked to building geometry as a single PMTiles object

File formats

The majority of the data is stored as Apache Parquet, with geospatial data being encoded as GeoParquet or as GPKG (GeoPackage) for direct download.

The data is split regionally based on NUTS1 regions.

  • {NUTS1}.parquet - GeoParquet 1.1.0, ZSTD compressed, with GeoArrow geometry encoding and covering bbox for fast spatial queries.
  • {NUTS1}.gpkg - GPKG (GeoPackage), direct download only.
  • {NUTS1}_morphotopes.parquet - GeoParquet 1.1.0, ZSTD compressed, with GeoArrow geometry encoding and covering bbox for fast spatial queries.
  • {NUTS1}_morphotopes.gpkg - GPKG (GeoPackage), direct download only.
  • {NUTS1}_data.parquet - Parquet, ZSTD compressed
  • morphotope_data.parquet - Parquet, Snappy compressed. All characters for non-outlier morphotopes - used to build the taxonomy.
  • label_name.json - JSON
  • pen_portraits.json - JSON
  • ward_linkage.npy - Numpy NPY
  • complete.pmtiles - PMTiles

Data repositories

Zenodo

The main data repository is Zenodo, which contains a complete copy of the data product and issues a DOI to each version.

2025-09 release

The initial release covering AT, CZ, DE, LT, PL, SK is avaialble from doi.org/10.5281/zenodo.17076283.

2025-11 release

The first expansion covering BE, NL, FR, ES is asvailable from doi.org/10.5281/zenodo.17600617.

Direct download

You can also download the data directly from the repository managed by Charles University.

2025-09 release

NUTS1 HiMoC on buildings HiMoC on morphotopes Morphometric data
CZ0 Parquet (423M)
GPKG (1.3G)		 Parquet (435M)
GPKG (930M)		 Parquet (1.7G)
NUTS1 HiMoC on buildings HiMoC on morphotopes Morphometric data
LT0 Parquet (188M)
GPKG (568M)		 Parquet (255M)
GPKG (580M)		 Parquet (332M)
NUTS1 HiMoC on buildings HiMoC on morphotopes Morphometric data
SK0 Parquet (275M)
GPKG (893M)		 Parquet (361M)
GPKG (759M)		 Parquet (1.1G)
File name Download
Morphotope data Parquet (200M)
Label to name mapping JSON (4K)
Descriptions of named branches JSON (8K)
Taxonomy linkage NPY (16M)
Complete classification as PMtiles PMTiles (7.8G)

2025-11 release

File name Download
Morphotope data Parquet (184M)
Complete classification as PMtiles PMTiles (16G)

License

This dataset is based on data from multiple sources. The combined database is licensed under the Open Database License (ODbL) v1.0.

Attribution: Contains data © Charles University, OpenStreetMap contributors, TomTom, GeoBasis-DE/LGB, GeoSN, LGL Baden-Württemberg, GeoBasis-DE/MV, GeoBasis-DE/LVGL-SL, IT.NRW, GeoBasis-DE/LGLN, GDI-Th, LVermGeo ST, GeoBasis-DE/LVermGeo SH, Bayerische Vermessungsverwaltung, LGV Hamburg, Landesamt GeoInformation Bremen, GUGiK (PL), ČÚZK (CZ), ÚGKK SR (SK), BEV (AT), Geoportal.lt (LT), IGN, 3DGI, DGC, and others.

Individual source building data is available under the following licenses and attributions: