I’m continuing on my quest to map all the bike racks in Chicago, but it’s going a bit slower than I thought. Last update, I had a little less than 200 racks mapped, which is pretty good progress. This update I’ve gotten to 306. Here’s a look:

I got a suggestion last time to try using data from the city to help map the racks. Unfortunately, the data from the city isn’t super accurate. It can be off by as much as 2 blocks using the map on the official website. So I got data that was a little more updated, stripped the locations from it (the geocoder used was inaccurate) and punched it into Google MyMaps to see how accurate it was. It’s better, but still not something I’d want to use for actual mapping.

Another approach I’ve tried is using Mapillary. I bought a 360 camera and recorded a few tracks around town. Supposedly they’ll automatically tag stuff using computer vision (bike racks included), but none of my sequences posted more than a week ago have been tagged yet.

Additionally, the camera I bought (GoPro Max) takes great photos, but it doesn’t have world lock or horizon leveling enabled for the timelapse feature that I use for Mapillary. So some of my photos look pretty off-kilter. But they’re still super helpful for mapping from home so I’m not having to physically go to all the locations I want to map!

My current approach is to map racks while I’m out at work (as you can see in the loop) or to map them using Mapillary data as I’ve ridden by (the ones in Lakeview and at Southport). This seems to work okay, but given the sheer amount of bike racks in Chicago I don’t know if it’s super scalable. I’m going to hope Mapillary figures out stuff on their end and auto-tags my tracks as that would make the process significantly faster.