Note – originally published over here on the CHI blog.
OGC is a project to build open data for addresses, turning a string like “12 Main Street” in to the latitude and longitude of where it is. The desktop site has two main pieces right now. First a simple way to explore the data:
Green pins have been verified, red are source data. You can turn pins green by surveying with the mobile app. Second, it has a number of simple online games which show you a pin and then ask whether it is reasonable to move it to the right place. These are a work in progress.
If you want to help fix the data, just open up the mobile site on your phone and walk around. Tap on the address where you are, and if it’s missing just add it. Simple:
The data is downloadable and contains 92 million points or so, exclusively in the US for now. Although, the project does work globally.
Why do this? Well, we need address data. With data like this, OpenStreetMap becomes usable for various end-user scenarios like in-car navigation. Today, to make this possible people typically license that data or infringe it for use with OSM.
Fundamental problems exist with OpenStreetMap itself collecting address data, and in fact very little exists there. There are other projects but they don’t have a consistent (or any) license and tend to rely on governments or VC/corporate money for support which could disappear any time.
Sustainable Open Source
To me, the question is how to build datasets where everyone wins and do it sustainably. We want this data to be open, but it’s kind of boring to ask volunteers to collect it, compared to making beautiful maps. On the other hand, there are plenty of companies that would be happy to pay for data like this if it was reasonably priced. Can we marry these two things?
In the software world, the answer is dual licensing. That’s the idea we’re exploring here with OpenGeoCodes. Private funding exists to pay people to collect this data in certain places and you’ll see us do that. With money, we’re able to pay people to fix the code and the data. We’re able to pay people to import new datasets and merge them, rather than waiting for volunteers who may not be interested. We’re able to work with you if you have updates to the data, to incorporate it, without the friction that a typical open project would entail.
And yet we can still make the data open. For now it’s CC-By-NC but this will likely become more liberal and additionally older data will be released under more open licenses. So you’ll be able to download old data under the public domain, where newer data may have a few restrictions or cost some money. Then we take that money and use it to improve the data for everyone.
OpenStreetMap will likely remain ODbL forever. If you think about it, this means the public domain mapping data isn’t improving at all. It would be nice if old OSM data dropped in the public domain after a year or two but this is fraught with difficulty. Here with OpenGeoCodes we can experiment with that and see if we can find some balance where we use money to improve this address data in partnership with open communities so that everyone wins.
Get in Touch
Have a need for geocoding data? Have data you can contribute back? Tired of expensive or badly licensed data? It’s very early days for this project. There are a number of tools and pieces of data coming down the road, and if you want to get involved please say hello or join the mailing list.