Overture Maps: where the businesses we count come from

What Overture Maps is

Overture Maps is a project of the Overture Maps Foundation, founded in December 2022 by Amazon Web Services, Meta, Microsoft and TomTom under the Linux Foundation, and since joined by many other members. It publishes open, free map data, ready to use in a map, an app, or an analysis like ours.

The data is split into themes:

• Places: businesses and points of interest, with name, category, address, phone and website.
• Buildings: the footprint, and often the height, of buildings.
• Transportation: roads, paths and railways.
• Divisions: administrative boundaries, from municipalities to countries.
• Addresses and base: street numbers, land, water and land use.

Why it exists

A reliable map is very expensive to build and to keep up to date, and for years every company redid the same work on its own. Overture exists to share that work: an open base of data, with one documented schema, that anyone can use and that companies can build their own services on.

Its stated goal is to merge several sources into one dataset, check its quality, and publish it in regular releases, so that anyone using it always knows which version they have.

Where the data comes from

Overture does not collect data in the field: it merges different sources, compares them and removes duplicates. Its sources include OpenStreetMap, data contributed by its members (for places mostly Meta and Microsoft, for roads also TomTom), buildings derived from satellite imagery, and open government datasets, such as addresses.

Each place carries the sources it comes from and a confidence score, which says how likely it is that the business really exists there. The data is published under open licenses (CDLA Permissive 2.0 for places, ODbL for the themes that contain OpenStreetMap data) and can be downloaded freely.

Our first attempt: OpenStreetMap

We built our first scans on OpenStreetMap, the free map maintained by volunteers around the world. For roads, paths, buildings and navigation it is excellent, often more detailed than commercial maps.

For businesses, though, the data goes stale. A shop that closed years ago stays on the map until a local volunteer notices, and the same goes for opening hours, phone numbers and above all website addresses: on many entries the website was gone, had changed, or had never been added. For a scan that starts from the website each business lists, that was the weakest point.

Overture handles this problem much better, because it merges several sources, including the listings businesses keep up to date themselves on its members' platforms, and compares them. That is why, since then, the state of the web scans count businesses from Overture Maps.

How we use it in the scans

For each city we take the administrative boundaries from the divisions theme and the businesses from the places theme: we count the ones whose point falls within the boundaries, after leaving out the categories that are not relevant to an online presence. Each business's website is the one listed on its Overture entry, and every check starts from there.

It remains a list of businesses found on a map, not the official business register: a business without an entry does not appear, and an entry with an old website leads to a site that does not open. That is also why the reports tell the businesses that list a website apart from those that have a working one.