Backed by Microsoft, AWS and Meta, the Overture Maps Foundation launches its first open map data sets

It's not often you'll find Microsoft, Amazon and Meta in the same room, collaborating toward the same goals. But that's exactly what we have with the Overture Maps Foundation, an initiative to develop interoperable and open map data.

Launched in December 2022, the Overture Maps Foundation is an attempt to counter Google's stranglehold on online mapping. The Linux Foundation-hosted outfit has been releasing early previews of its data sets over the past year, and the first beta incarnation arrived this April. But on Wednesday, we're seeing the first formal fruits of this big collaborative effort: The initiative is launching a quartet of open data sets in general availability (GA).

For context, maps are essentially "layers" that can be tailored to many uses. The Overture Maps Foundation is today releasing buildings, constituting 2.3 billion building "footprints" globally; places, which includes some 54 million notable places of interest; divisions, which serves as a visual overlay denoting "boundaries" separating countries, regions, cities or neighborhoods; and base, which covers land and water features such as physical infrastructure (e.g. communication towers, piers and bridges).

The company's other main dataset, transportation, will remain in beta for now. It is also debuting a new addresses dataset in alpha, which supports 200 million addresses across 14 countries.

There's a map for that

While Microsoft, AWS and Meta are the highest-profile members of the Overture Maps Foundation, the core steering committee also counts location technology stalwart, TomTom, as a member. Other "general" and "contributor" members include Esri, Hynundai, Niantic and TripAdvisor. Google's absence from the group is notable, albeit unsurprising given the ultimate goal of the project.

Collectively, the members are pooling myriad data sources, including open datasets from tangential projects such as OpenStreetMap and government sources, their own internal proprietary data, and even data from the main nemesis here, Google.

They can do that because although Google's mapping data empire is mostly proprietary, it has released some datasets under an open access license, including Open Buildings, released back in 2021. As we can see from this map of the U.S. / Mexico border, the Overture Maps Foundation has used data from OpenStreetMap, Esri, Microsoft and Google.

Building data along the U.S/.Mexican border, collated from multiple sources
Building data along the U.S/.Mexican border, collated from multiple sources

It may sound simple to combine data sets, but the reality is somewhat different, as they generally don't adhere to the same formats, structures and standards. So you might have two largely similar data sets with slightly different purposes that need to be meshed together to integrate their respective benefits. The process of bringing together such data sets is called conflation, and it can be a painstaking process of checking and de-duplication.

"One of the real challenges when you start combining data that's coming from a lot of different places is, how do you know that this record of a building or an address or a place is the same as this other record?" explained Marc Prioleau, executive director of Overture Maps Foundation, in an interview with TechCrunch. "That seems kind of obvious, but people misspell things or use different names. They could also be slightly misaligned geographically. Conflation plays a big part in [fixing] this."

Community mapping project OpenStreetMap is playing a huge part in providing data for the Overture Maps Foundation, as do open data sets provided by governments and municipalities. But the third major category particularly notable here is that of the corporate members themselves, which are pooling what they have in-house.

"The best example of this is that Meta took its [Facebook] places data — as you can imagine, every 'place' or restaurant has a Facebook page — and shared it with the project," Prioleau said. "Facebook made an interesting decision there. While it has a competitive advantage in knowing what people like about a place, what they serve, all those pictures and social media posts, simply knowing that the place is there is not a competitive advantage. But if they shared that, and got other companies to share similar data, then we would have the advantage of having a more complete record."

Prioleau himself has a fairly long history in the mapping and location realm, having worked in related roles at Mapbox, Uber and more recently as head of business development for mapping and location at Meta, a position he left last May to lead the Overture Maps Foundation. He was joined by Amy Rose as technical director in November, meaning the organization has just two full-time employees.

On top of that, the foundation counts around 10 contractors who do everything from marketing to project management, and the host, Linux Foundation, also contributes some engineering support. This is where the member companies enter the fray, contributing around 100 people across engineering and product management, and who might help out some of the time or all of the time, depending on the task at hand.

"When our companies [members] come in, we really emphasize that this is not a spectator sport — it's a participation sport," Prioleau said. "So we ask people to go and work on one of the many pieces of the project."