Building models of running routes

On the idea of creating routing algorithms to discover new places

September 10, 2025

Is running simply about following edges on a graph? This is what I was wondering on one of my recent runs, having worked with OpenStreetMap data to find roads that are free of traffic. I tried to write an algorithm that would generate a route for running based on my preferences. The routes I got back were very machine-like. There is precision but not the ability to make the kind of turns I would make on my daily runs. Perhaps the problem is that I can never have enough data points to recreate the decisions I make to take this path or that. What I also realized is that a computer version of the routing network is incomplete. This is not the kind of problem that can be solved by mapping more streets. I believe it is not possible to document every single path available because there is no need to do so at present. It is good enough. There is also a distinct problem of perception—of what you see from street level. No amount of looking at the map will give you that sense. A 3D model is painfully hard to create by hand, and it does not feel adequate to the reality you experience when you run. The lived experience is infinitely richer and has more dimensions. Enjoy your runs. They are truly yours.