More contributions through specialized GUIs

The success of GalaxyZoo in crowdsourcing huge amounts of data has been on my mind a bit, and I believe OpenStreetMap could learn a thing or two from their approach.

There are a few big differences between GZ and OSM contributions. As I see it the most important are:

  1. GZ has a simpler dataset to collect: spiral/globular/edge-on/barred, direction of spin. This compared to OSM where you have to place a lot of points/lines AND classify them and the classifications are pretty vast.

  2. OSM presents a huge canvas and basically says “go for it” while GZ presents just a single task to accomplish at a time. This makes OSM vulnerable to something similar to writers block where it’s hard to take a dicision on where to start.

  3. The OSM editor is very big and can do a lot, while GZ has a nice decision tree of most common options with just a few options at each step, and one fallback “this is something else” option.

I believe OSM can start to collect much more data rapidly with some specialized user interfaces. My idea is to have three specialized apps:

  1. Missing buildings: Click on buildings that haven’t been drawn
    To produce good areas for this app to show to users there could be a database of areas with poor mapping coverage. This could either be supplied by yet another specialized app or by some combination of manual work and algorithms to find city blocks by analyzing roads.

  2. Draw a building: You are shown a single building and you draw it
    The previous app would obviously supply the dataset for this app. The good thing about this app is that just drawing a building and then getting another one to draw is that you don’t ever see how many buildings are yet to be completed, avoiding the sense of futility in the face of large numbers

  3. Fix bugs: draw a random bug from the bug pile, and fix it
    Pretty straight forward I think, and could easily be implemented on top of the current editor.

All these apps would keep track of how many tasks you finished in each category. This would produce a reward/reputation system similar to for example StackOverflow. One could also add some facebook integration to post achievements etc to try to push friends to keep up with your pace.

Let me know what you think!

Great ideas! I’ve had similar ideas myself, but haven’t had the time. These would be great Google Summer of Code projects.

I see at least two other major aspects of OSM you have not mentioned, and that are relevant for how a specialized mapping tool should be built to work well within the OSM community.

  1. At OSM, we prefer local knowledge. Mapping places you have never been to in real life does happen, but is often considered a questionable practice, as it produces poor data quality: This is due to outdated aerial imagery, incorrect interpretation of said imagery, and because many very important facts (road names, access limitations, house numbers, …) cannot even possibly be derived from aerial imagery. Obviously, having been to that place is not generally expected for GZ contributors. :wink:

This doesn’t preclude the possibility of specialized apps, but they should at least let the user choose the area to work in. Alternatively, they could be location-aware and encourage outdoor activity. Mobile devices might actually be well suited as a platform.

  1. We ideally want long-term data maintainers. If someone just maps a few buildings, that’s of course fine. But it’s much more valuable to have contributors who feel responsible for updating it. Things like reading about construction activity in the local newspaper and instantly knowing what needs to be changed in OSM, because they know what the OSM data currently looks like in that place.

The issue with maintenance is that there will probably not be any meaningful metrics for completion of this ongoing task. So this challenge is, imo, badly suited for any points-based gamification. The same is true for other crucial tasks, and it doesn’t seem to send the right message when you get “points” for low-priority tasks such as tracing building outlines, but none for more important activities.

I agree that local knowledge is preferable but I don’t see it as a necessity. I believe it’s much easier to update incorrect data than to add it all from scratch. Like you say adding house numbers on buildings would be much easier if you could just add house number and not have to also draw every single house (which you probably need to do based on satellite data anyway for practical reasons). That’s also why Mapzen POI Collector is a great complement to the OSM website.

Do we really? I don’t think you do. You want a few long-term maintainers, but if the vast majority of drudgery is offloaded from these highly interested and motivated few people they can get much more important work done like the examples you gave.

You get points for the amount of times you click on fields in FarmVille. That is by definition 100% useless work, but you still get points from it. I don’t see why this would make taking out the trash or going to the grocery store seem less important :stuck_out_tongue:

My idea is not based on trying to gamify OSM any more than GalaxyZoo is trying to gamify astronomy. No astronomer will say to himself “screw this, I don’t get any points for finding stuff out about the universe but I do get points on GalaxyZoo, I’ll just do that instead”. What GZ does do is enable types and scales of data acquisition that was impossible before. GalaxyZoo doesn’t diminish astronomy, and an OSMZoo wouldn’t diminish OSM.

Also, how do we get these long-term maintainers on board? I believe the first order of business has to be to make people aware that OSM exists and understand what it is. The apps I suggested above could create a lot of buzz which could in turn enable a lot of people who are interested in maps to discover that OSM is actually their life long passion, they just didn’t know about it.

There are lots of people with great ideas like this. We do not have a shortage of good ideas.

Unfortunately, great ideas don’t really get much done in themselves, and we do have a shortage of developers willing to work on the core tools.

So if you’d like to help make OSM more user-friendly, roll up your coding sleeves and get involved.