Geo Week News spoke with Executive Director Marc Prioleau about the complexity around this release.
Towards the end of 2022, the Linux Foundation announced the formation of a new group called Overture Maps Foundation, comprised of founding members including Meta, Amazon Web Services (AWS), Microsoft, and TomTom, dedicated to developing interoperable open map data. It was an ambitious goal, but the group has moved quickly since that formation, in many ways culminating this summer when they announced the general availability of most of their open map datasets, something Geo Week News spoke with Overture’s Executive Director, Marc Prioleau, about at the time.
Within that announcement was a note that the one dataset that was not available at that time was their transportation dataset, which was a bit more complex and thus lagged behind a bit. Today, the organization announced the general availability of this global transportation dataset. In their announcement, they say the dataset “supports new and expanded use cases across a broad swath of industries including automotive, ride-sharing, logistics, navigation, local search, urban planning, and disaster and humanitarian response.”
As you can see in the image above showing a heat map of this transportation dataset’s coverage, it is quite comprehensive. Overture notes that the release includes 86 million kilometers of roads around the world and that the data has been in use by early adopters like Microsoft, Meta, and TomTom. Now, other developers will be able to start using this data, joining the rest of the datasets that were part of the previous general release, including buildings, places of interest, divisions, and a base layer.
In a recent conversation with Geo Week News, Prioleau dove a bit deeper into what exactly made this transportation layer more complex. Most simply, he noted that the hard part of this dataset was the “interconnectedness of it.” He points out that the other datasets released in July were largely standard and static – an address is a point on the map, a building footprint is a polygon, etc. For roads, however, it’s a more connected and flowing dataset that can be hard to make as a standard dataset.
“If I want an identifier for this building, that’s fairly easy,” Prioleau said. “But a road ends up being a road segment, and things happen along those segments. Traffic happens along it, potholes happen along it, police cars might be located along it. You end up having more of a linear reference system, which is a more complex piece to it.”
He goes on to explain that the majority of this dataset is from OpenStreetMap, which is a crowd-sourced dataset. This is great in getting data from disparate parts of the globe, but it also means that the data can be inconsistent. As Prioleau points out, people map differently.
“If you look at [OpenStreetMap data] visually, you might go, Wow, that looks really good. If you looked at it with kind of X-ray vision, what you would see is those lines and curves are all drawn by people going in and mapping them. People map differently. I might put a ndoe here and a node five miles away if it’s a straight line, but you might break it up differently. If you think about attaching data to it, that needs to be a logical segment.”
That, in essence, is where the complexity and eventual delay in this release comes from. Overture had to figure out a way to standardize these crowd-sourced maps in a logical way. This required a lot of work from the developer team working with Overture.
“Essentially, what we had to do is come up with rules,” Prioleau explained. “What is a road segment? An obvious example might be from an intersection to another intersection; the strip between is a road segment.”
Overture set up a number of rules to “define” these road segments, including things like intersections as well as changes in the number of lanes, changes in speed limit, etc. Once those rules were defined, they could set up an automated way to reformat and redo the data, ultimately leading to being able to release this dataset widely.
This release marks another major milestone for Overture, and puts them on a path for more goals and work in the coming year. Among the next things on the checklist, according to Prioleau, is working on a way to validate the quality of their data. First, they had to get the data out there, but now it’s a matter of going over with a fine-toothed comb to validate. He also mentioned working on their global entity reference system, or GERS. And finally, they can start to concentrate more on greater adoption. We saw in the 30-Day Map Challenge that Overture is getting some usage, and the organization is looking to grow that adoption even more in the coming year.
