Hugging Face's logo

Datasets:
wikimedia
/
linked-openstreetmap-wikidata-wikipedia-wikivoyage

Tasks:
Text Retrieval
Modalities:
Tabular
Text
Geospatial
Formats:
parquet
Languages:
English
Size:
10M - 100M
Tags:
openstreetmap
wikidata
wikipedia
wikivoyage
geospatial
geoparquet
Libraries:
Datasets
pandas
Polars
License:
Dataset card Data Studio Files
xet
 Community
Dataset Viewer
Auto-converted to Parquet Duplicate

The dataset viewer should be available soon. Please retry later.

Linked OpenStreetMap × Wikidata × Wikipedia × Wikivoyage

Users of open geospatial data frequently reach for data from OpenStreetMap and Wikimedia solve different problems. OSM includes high-resolution global map data of roads, buildings, natural features, and political divisions. Wikimedia offers narrative, historical, and semantic coverage of these same entities via projects like Wikipedia, Wikivoyage, and Wikidata. Both are rich with foreign key references to the other, but neither has yet published a fully expanded linked dataset joining the two sides.

This release selects Wikimedia items with spatial properties, OpenStreetMap elements with Wikimedia references, and offers a complete linking table between the two reflecting their complete global status as of September 15, 2026.

A central table links these data sources together in one reference lookup table, links.parquet. This is the core of the dataset to use for disambiguation or interlinking. The original data sources, extracted at the same time the central table was constructed, are also included for completeness.

Dataset Summary

Excerpt of Wikipedia, Wikidata, Wikivoyage, and OpenStreetMap datasets with all geographic information linked between the four:

  • 17m OpenStreetMap elements
  • 13m Wikidata items
  • 1.5m Wikipedia articles
  • 31k Wikivoyage articles

Elements, items, and articles are included if they have or link to geographic information like Wikidata lat/lon coordinates or OpenStreetMap element IDs. The dataset is provided in Parquet format, optimized for high-performance analytical queries and efficient storage. This version uses a unified, pinned schema across all files, making it compatible with DuckDB, pandas, Polars, and Apache Spark out of the box.

Where applicable, includes these external properties:

  • Wikipedia and Wikivoyage article content
  • OpenStreetMap node, way, and relation elements connected to the item
  • Geoparquet representation of OSM elements connected to the item
  • QRank value indicating global significance of the item based on external signals (https://qrank.toolforge.org)

Attribution & License

This dataset is composed of four upstream open datasets, Wikidata, OpenStreetMap, Wikipedia, and Wikivoyage, each created and maintained by international volunteer communities. The dataset contains six files licensed under the terms of their respective source data:

  • Links.parquet, Wikidata.parquet, and Wikidata_locations.parquet are made available under the terms of Creative Commons CC0 License
  • OpenStreetMap.parquet is made available under the terms of the Open Database License
  • Wikipedia.parquet and Wikivoyage.parquet are made available under the terms of the Creative Commons CC-BY-SA Licence

When publicly using the OpenStreetMap portions of this dataset, you are required to do the following two things:

  • Provide credit to OpenStreetMap by displaying an OSM attribution notice
  • Make clear that the data is available under the Open Database License

When publicly using the CC-BY-SA portions of this dataset, you are required to incorporate this text:

OpenStreetMap

OpenStreetMap provides map data for thousands of websites, mobile apps, and hardware devices. OpenStreetMap is built by a community of mappers that contribute and maintain data about roads, trails, cafés, railway stations, and much more, all over the world.

Wikidata

Wikidata is a free and open knowledge base that can be read and edited by both humans and machines. Wikidata acts as central storage for the structured data of its Wikimedia sister projects including Wikipedia, Wikivoyage, Wiktionary, Wikisource, and others.

Wikipedia

Wikivoyage

Usage Examples

Parquet files can be queried with SQL queries using tools like DuckDB.

Find the most significant entities inside the Quartier du Gros-Caillou, according to QRank:

WITH box AS (  
  SELECT bbox  
  FROM read_parquet('openstreetmap.parquet')  
  WHERE osm_type = 'relation'  
    AND tags['wikidata'] = 'Q3413275' -- Quartier du Gros-Caillou  
  LIMIT 1  
)  
SELECT w.label, w.qrank, o.osm_type, o.osm_id, o.url  
FROM read_parquet('openstreetmap.parquet') o, box  
JOIN read_parquet('wikidata.parquet') w  
  ON w.wikidata_id = o.tags['wikidata']  
WHERE o.tags['wikidata'] IS NOT NULL  
  AND o.bbox.xmin >= box.bbox.xmin AND o.bbox.xmax <= box.bbox.xmax  
  AND o.bbox.ymin >= box.bbox.ymin AND o.bbox.ymax <= box.bbox.ymax  
  AND w.qrank IS NOT NULL  
ORDER BY w.qrank DESC  
LIMIT 10

Results:

Eiffel Tower https://www.openstreetmap.org/way/5013364 9,186,031
Organisation internationale de la Francophonie https://www.openstreetmap.org/node/2863826619 581,838
Champ de Mars https://www.openstreetmap.org/way/4208595 339,745
Tunisia campaign https://www.openstreetmap.org/way/427592604 276,138
Musée du quai Branly https://www.openstreetmap.org/way/17954721 137,339

Find the administrative regions enclosing the Eiffel Tower:

WITH anchor AS ( SELECT 'Eiffel Tower' AS title ),  
anchor_page AS (  
  SELECT wp.page_id, wp.wikidata_id  
  FROM read_parquet('wikipedia.parquet') wp  
  JOIN anchor ON wp.title = anchor.title  
),  
anchor_pt AS (  
  SELECT wl.geometry AS g, wl.longitude AS lon, wl.latitude AS lat  
  FROM read_parquet('wikidata_locations.parquet') wl  
  JOIN anchor_page ap ON wl.wikidata_id = ap.wikidata_id  
  ORDER BY wl.precision NULLS LAST  
  LIMIT 1  
),  
regions AS (  
  SELECT o.osm_id,  
         o.tags['admin_level'] AS admin_level,  
         o.tags['name'] AS name,  
         O.tags['wikidata'] AS wd  
  FROM read_parquet('openstreetmap.parquet') o, anchor_pt a  
  WHERE o.osm_type = 'relation'  
    AND o.tags['boundary'] = 'administrative'  
    AND o.bbox.xmin <= a.lon AND o.bbox.xmax >= a.lon  
    AND o.bbox.ymin <= a.lat AND o.bbox.ymax >= a.lat  
    AND ST_Contains(o.geometry, a.g)  
)  
SELECT r.admin_level, r.name, r.wd, r.osm_id, lk.wikivoyage_page_id AS wv_page  
FROM regions r  
LEFT JOIN read_parquet('links.parquet') lk  
  ON lk.wikidata_id = r.wd  
 AND lk.source = 'wikidata_sitelink_wikivoyage'  
ORDER BY CAST(r.admin_level AS INTEGER) DESC

Dataset Structure

Links

Links.parquet supplies 1:1 connections between the other five tables via these columns:

  • Source – string describing the source of the link, e.g. “osm_wikidata” for a QID reference from OpenStreetMap to Wikidata or “wikidata_p402” for an relation ID reference from Wikidata to OSM
  • Wikidata_id – Wikidata item QID
  • OSM_type – OSM element type when present
  • OSM_id – OSM element ID
  • OSM_tag – string key of the OSM tag referencing an item or article when present
  • Wikipedia_page_id – numeric page ID of Wikipedia article when present
  • Wikivoyage_page_id – numeric page ID of Wikivoyage article when present

We include 24m links in a 174MB parquet file.

OpenStreetMap

OpenStreetMap.parquet stores all OSM elements in these columns:

  • OSM_type – OSM element type
  • OSM_id – OSM element ID
  • URL – web address of element
  • Version – numeric version of element
  • Timestamp – last-edited timestamp for element
  • Tags – map of string/string freeform tags
  • Geometry – unprojected Geoparquet geometry
  • BBox – struct with xmin, ymin, xmax, and ymax for lat/lon envelope
  • Elements – JSON blob of OSM sub-elements comprising this one

We include 17m OSM elements in a 36GB geoparquet file.

Wikidata

Wikidata.parquet stores Wikidata items:

  • Wikidata_id – item QID
  • URL – web address of item
  • Label – items’s English label
  • Description – item’s English description
  • Instance_id – array of P31 QIDs describing item’s type
  • Modified – last-modified timestamp
  • QRank – global popularity score (higher = more prominent)

Wikidata_locations.parquet stores Wikidata locations when these are simple lat/lon coords:

  • Wikidata_id – item QID
  • Qualifier_property, qualifier_label, qualifier_value – type of location
  • Latitude, longitude – unprojected degree coordinates
  • Precision – coordinate precision ranging up to 1.0 degrees for coarsest precision
  • Altitude – numeric altitude when given
  • Geometry – unprojected geoparquet point
  • BBox –  struct with xmin, ymin, xmax, and ymax for lat/lon envelope with precision

We include 13m Wikidata items in a 436MB file, and 13m locations in a 957MB file.

Includes Wikidata items with any combination of these location properties:

Additionally includes Wikidata items referenced via “Q” IDs from OpenStreetMap:

Wikipedia & Wikivoyage

Wikipedia.parquet and wikivoyage.parquet share these core columns:

  • Page_id – numeric page ID
  • URL – web address of article
  • Title – primary title of article
  • Abstract – lead summary text
  • Date_modified – last-edit timestamp
  • Wikidata_id – main QID link from article to Wikidata, when present
  • Article_body_html – fully rendered article HTML
  • Article_body_wikitext – raw wikitext source

We include 1.5m Wikipedia articles in a 22GB file and 31k Wikivoyage articles in a 515MB file.

Invitation For Feedback

We welcome feedback to help refine and improve this dataset. Please share suggestions or issues on the Wikimedia Enterprise Meta-Wiki discussion page.

Downloads last month
99
Number of rows:
50,704,133
Total file size:
59.4 GB