World-scale OpenStreetMap POIs in CSV

After multiple requests from POI value chain stakeholders, we are happy to announce that SLIPO now provides a free download service for global POIs from the OpenStreetMap (OSM) crowdsourced database in CSV (comma separated values) format.

This is a follow-up to our existing service that we launched a year ago and provides monthly dumps of POIs in RDF triples as extracted from OSM, one of the most valuable global geospatial repositories for the industry, innovators and researchers.

Extracting this wealth of POI data into a neutral tabular format can greatly facilitate loading of the file contents in DBMSs and GIS for further analysis, since CSV is widely used for data exchange in the industry.

The extracted POI data is organized in zip archives of CSV files per country or region and contain more than 18.5 million POIs worldwide. This archive is updated once a month, so that you can download the latest POI data for your line of work.

We extract POIs from the original OpenStreetMap data using our OSMWrangle ETL software developed in the context of SLIPO and SmartDataLake projects. This generic-purpose, open-source tool can be used for extracting features from OpenStreetMap files and transforming them into records in a delimited CSV file. OSMWrangle draws much of its functionality from the source code of TripleGeo, our Extract-Transform-Load tool into RDF developed in the context of SLIPO.

OMSWrangle applies a custom filtering against the original OpenStreetMap data in order to select features that represent Points of Interest (POIs) ranging from bars, restaurants, and museums to ATMS or bus stops from the rest (e.g., road networks, rivers, administrative boundaries etc.).

All POI features retain their original OpenStreetMap identifiers. Certain tagged values of the original OSM features are extracted in specific attributes as denoted in the header of each CSV file:

  • Detailed geometry (WKT) in WGS84;
  • Lon/lat coordinates for the centroid (in WGS84);
  • Name (usually in local language);
  • International name (usually in English or transliterated in Latin characters);
  • Alternative name;
  • POI classification (in 14 categories and 167 subcategories) based on OSM tags;
  • Address (street, house number, postcode, city, country);
  • Contact (phone number, fax number, website, operator);
  • Opening hours;
  • Full description (if listed in OSM);
  • Link to a photo/image;
  • Link to Wikipedia article;
  • Date of last update,

while the remaining tagged values are listed under the ‘OTHER_TAGS’ column.

Geometries of OpenStreetMap POIs are represented as Well-Known Text (WKT) literals according to the OGC standard. Original OpenStreetMap nodes, ways or relations are suitably converted into WKT geometries, i.e., (multi)points, (multi)linestrings, (multi)polygons, geometry collections, etc.

Feel free to visit our free service and download POIs in CSV format for any part of the world!