KITT Come Get Me! Mapping Parking Lots in OpenStreetMap with MapRoulette

Tesla rolled out their v10 car software late September. A notable and much discussed feature it introduced is Smart Summon. Your Tesla will come find you in a parking lot.

Tesla owners were quick to figure out that this works better in some lots than in others, and the reason why is better mapping of roads within parking lots.

Better mapped parking lots benefit not only Tesla but also all other users of OpenStreetMap data, and they make the map look better too.

It can be tricky to know where to find parking lots in OSM that lack detail. This is why I created a MapRoulette Challenge to serve them to aspiring lot mappers one by one. If you’re unfamiliar, MapRoulette is the microtasking platform for OpenStreetMap I came up with in 2012 and have been building out, with help of a great team, since then.

If you haven’t used MapRoulette before, here’s a quick step by step:

  • Go to the Challenge page on MapRoulette
  • Click ‘Sign in to get started’. This will take you to OSM to sign in
  • After you’re redirected to MapRoulette, click ‘Start’
  • You are served a random Task. Click ‘Edit’ to edit the area displayed on the map in OSM
  • Improve the parking lot by adding parking aisles. See the Task instructions for details
  • Upload the changes to OSM
  • Return to the MapRoulette tab in your browser
  • Confirm that you completed the task
  • Repeat as many times as you like

Are you interested in how to create Challenges like this? I am writing a post detailing step by step how I created this one. Stay tuned for that, or head over to the MapRoulette documentation right now for general instructions.

TIGER and OSM, 12 years later

More than a million people have edited OpenStreetMap (OSM) for the last 15+ years (and I am one of them). That means we have a map that is fantastic to look at in many places, like here at Epcot Center in Florida, USA.

OSM mappers love to highlight great and impactful mapping accomplishments like Map Kibera and the massive campaign followingthe devastating Typhoon Haiyan. And rightfully so, OSM shines where no other map can because it is open and free.

There is also a shadow side to OSM, areas that nobody seems to care about. You could argue that this doesn’t matter. OSM is, in a way, subject to the laws of supply and demand. Where a better map is needed, one shall be created.

But OSM has never really been about that. Most volunteer OSM mappers create the map with highly altruistic motivations and a sense of “personal but shared need”. Because of this, I don’t believe it’s too much of a stretch to posit that Linus’s Law — “Given enough eyeballs, all bugs are shallow” — applies to OSM as well. If you point out where OSM could use some love, mappers will extend that love and make the map better.

My deep faith in that power of a map-loving OSM community has inspired me to build several tools over the years aimed at guiding mappers to forgotten places in OSM. The best known is probably MapRoulette, a website that dispenses micro-tasks for mappers to solve.

A lesser known tool I built some years ago when I was at Telenav looking into the quality of the OSM road network in the USA is the TIGER Battlegrid.

The original TIGER Battlegrid, screenshot from the OSM wiki

The TIGER Battlegrid would show an OSM map overlaid with colored cells. The brighter the cell, the worse the quality of road alignment. It was called the TIGER Battlegrid because 1) The quality analysis was done by comparing OSM data with the then-current 2012 Census TIGER road network data, and 2) because OpenStreetMap in the United States got started by a huge import of TIGER road network data in 2007–8.

Animation showing TIGER data import, from the OSM wiki

The quality of that 2007 TIGER data was very poor in places and a lot of fixup was needed, something the community picked up quickly in many places. But the United States is a very big place and many areas remained where no fixup was done. So I cooked up the TIGER Battlegrid to point mappers to areas that still needed fixing.

The Battlegrid was popular and many places received some much needed mapping love. But I moved on to other projects, and at some point the Battlegrid went offline and, due to me neglecting to properly document the project, was lost.

People kept asking me about it and it made me sad to have to tell them no. Even though many data layers and tools are now available to inspect and detect outdated roads in OSM (see iD screenshot below), the Battlegrid struck a chord with people.

The OSM iD Editor with a colored overlay showing where mismatches exist between current TIGER roads and OSM.

And we are still far from done cleaning up after the 2007–8 TIGER import, as this image recently posted to the OSM Slack channel demonstrates:

Bad TIGER roads. Screenshot uploaded by Richard Welty, Road data ©OSM, aerial imagery ©Bing

So, I spent a bit of time recreating something close to the original Battlegrid, and you can preview it here. It only covers Colorado for now, it lacks some of the functionality of the original (notably, you can’t mark cells as done). And last but not least, the project in its current state is extremely hacky — but at least, it’s documented.

The current Colorado Battlegrid

Please give it a spin and let me know your feedback. I would love your help making it better.