Check The Places

Check places in OpenStreetMap against data from All The Places

About

OpenStreetMap is a free, community-made map of the world that anyone can edit and use.

All The Places is an open‑source collection of web scrapers to extract geographic data about businesses such as shop or fast food chains.

Check The Places compares data from All The Places with OpenStreetMap to help mappers keep OpenStreetMap data complete and up-to-date.

Get started

🇩🇪 Germany

Type ATP count OSM count Unmatched ATP Unmatched OSM Duplicates OSM
Shop 27107 26724 2226 1829 57
Amenity 7121 6641 850 366 48

How it works

The comparison between All The Places and OpenStreetMap happens by brand and by country. All The Places runs their scrapers and publishes the results roughly every two weeks. The data for each brand is downloaded from the latest run.

OpenStreetMap is queried via the Overpass API. When querying for one chain such as “Netto” supermarkets in Germany, the query would be for the expected main type, shop=supermarket in this case, for the name to be Netto Marken-Discount or Netto City, or for the brand:wikidata to be Q879858.

The assumptions are that the brand websites are more up-to-date than OpenStreetMap regarding the existence or non-existence of their branches. OpenStreetMap typically lags behind store openings or closures. On the other hand, the geocoordinates in OpenStreetMap are very precise, while the brand websites’ often are not. Many use geocoded addresses which can be off quite a bit.

An attempt is made to find a one-to-one match between the two location data sets. First they are matched based on the website. This works well if each store has a unique URL. Next, stores are matched by spatial proximity. The OpenStreetMap data is also checked for duplicates, which can happen e.g. if a store is tagged on a building outline, but also as additional point inside the building.

image/svg+xml All The Places OpenStreetMap Matching between All The Places and OpenStreetMap Matched by website Matched by proximity Only in ATP Only in OSM Duplicate in OSM

We get four sets of points out of this process:

  1. Matched points. We can check All The Places’ data for additional info such as website or opening hours which could be added to OpenStreetMap.
  2. Unmatched in OpenStreetMap. These might be new stores which are not yet mapped, stores which are mapped with wrong name or brand tags, or many other such cases. A mapper should take a look and add the store to OpenStreetMap if it is indeed missing.
  3. Unmatched in All The Places. These might be stores that have been closed. A mapper should take a look, investigate the object’s history and delete it if the store indeed doesn’t exist any longer.
  4. Potential duplicates in OpenStreetMap. If they are indeed duplicates they should be merged. Sometimes two different stores are very close by, in which case they should be left alone.

No changes are made directly to OpenStreetMap. The output of the process are one report per brand, listing above findings. An OpenStreetMap mapper can take these as hints which locations to investigate and decide on their own if and how to improve the data.

How to contribute

Browse the reports, pick a brand and check the reported differences in location data. Edit OpenStreetMap to add missing shops or to update outdated information.

If there‘s a brand you would like to see covered, please check All The Places whether they already have a spider for it. If not, please see their documentation how to contribute.

If there’s a brand where a spider already exists, a new top-level tag in OpenStreetMap or a new country you would like to see covered, please create a ticket (TODO where?). New categories or countries can be added, but there should be people willing to work on them. By creating a ticket you can show that there’s demand.

Project-wide metrics

Metric Count
ATP count 34228
OSM count 33365
Unmatched ATP 3076
Unmatched OSM 2195
Duplicates OSM 105

See Also

Check out these similar projects:

License

This project’s source code is free to use and modify under the MIT license.

The data generated by the project which is derived from OpenStreetMap data is © OpenStreetMap contributors and licensed under Open Data Commons Open Database License (ODbL).