Today I would like to share what I have learned about using external data in OSM but not in the classic import fashion.

I have gone over what I have learned about data importing in OSM after being involved in a few of them. But today I want to talk about how we’re using data from our local government as a reference but not using the usual import method and why.

So what is the definition of an import in regards to OSM? The Wiki defines it as “the process of uploading external data to OSM.“.

But uploading is not always the most effective way to use external data to improve OSM. What we are doing is showing external data as a layer over OSM and using that to edit OSM by hand.

But Why?..Wouldn’t it be better to direct upload the data?

Not always. We have found that using other people’s data for OSM comes with a lot of hurdles and direct importing can do more damage than good, but that’s no reason to give up on using the data. We just change methods to make it work. As I said in my last post, it’s important to get governments and organizations contributing to OSM, so we want to find a way to make use of data that is not suitable for direct import into OSM. As you can see in the picture above there is a layer that we can use to trace over. But you also can see that the subdivision is split into 3 blocks. This is probably due to the fact that it was built in that order. In this data set we see this concept referred to as sections, blocks, phases, and numbers like 1,2,3 smaller sections inside of subdivisions like “commons” and “place”. Naming can be somewhat unpredictable. There are also other things in the data that aren’t great for us, like plots of land that have the names of the development company or name of the owner, things that are not ground truthable so they need to be left out. It’s important to keep in mind that the people collecting this data were not doing it for OSM, and it’s the responsibly of the people bringing in the data to bring it up to local OSM standards.

But couldn’t we figure out those variables, account for all that, then do a normal import?

Well going through all the trouble of merging and fixes errors in the data takes longer than just fixing it on the fly. Plus after we have finished doing it by hand, everything is to our local OSM standards, and we have a good place to measure against and track further progress as we automate the process. Also we find many places where we use different tags and draw the lines differently than the county GIS department, so doing it by hand lets us make those judgment calls on a per area basis.

So is it an import or not?

Strictly according to the definition on the Wiki, No I don’t think so. However were still using external data to improve OSM so it is important to get permission, document what you’re doing, track the work, and give attribution. Even though the actual edits are coming from our mouse clicks and not automated software. I do think that there should be direction on the wiki for this kind of “reference” work where you get many of the same prerequisites and plan of action to input the data as an import. Both to make sure the data going into OSM is legally allowed and to provide guidance on what tools and methods to use. I’ve seen other structured tasks like this called Map-A-Thons and campaigns on various sites around OSM. I think the term “Campaign” fits well, and having documentation and tooling and best practices around this concept would be a good thing for mappers who want to make use of data that’s not in good shape for direct importing.

You can look at the code and follow along with the residential area import that were working on now at the links below.

https://github.com/OpenStreetMapSC/SpartanburgSubdivisions

https://tasks.openstreetmap.us/project/143

osm.wiki/Spartanburg_County_Subdivision_Import

osm.wiki/Import