I’ve been turning raw OSM signal nodes (highway=traffic_signals / crossing=traffic_signals) into one clean point per intersection, clustering the poles and clipping to the real city boundary. Covers 225 US and 271 EU cities. Vehicular counts come within ~4% of official city figures (LA 4,856 vs ~4,850, NYC 13,198 vs ~13,770).
One thing I’d like input on: when a junction has separate crossing=traffic_signals pedestrian nodes nearby, do you treat those as part of the same intersection or as distinct ones? Right now I fold them in if they’re within the cluster radius.
It’s a bit difficult to give feedback if you don’t include any links to your work. I would also suggest fixing the formatting on this post, as the way it is currently formatted makes it quite hard to read.
As a new user, you probably aren’t allowed to insert links in this post, but I don’t see any links in your forum or osm.org profiles either. Your findings are intriguing and encouraging; that’s why we’re interested in checking them out.
Thanks, that’s fair. It was a half-finished draft and the formatting came out messy. I’ve added the link to my profile now and put a cleaner version. Happy to share specific city examples if that’s useful.
Thanks Minh, you’re right, as a new user I couldn’t add links in the post and hadn’t filled in my profile yet either. Both sorted now. Appreciate the interest, would genuinely value your feedback. Thanks
So this is a paid service where I could pay you to give me a list of, lets say all signalled intersections in Berlin for the low low price of 63 euro’s?
I’ll be quite blunt and say that I’m not sure whether this is a genuine attempt at getting real feedback, or simply an attempt at advertising your business.
Edit: Okay no, you are part of the problem if you use a website like this: