Hi, I have been using cycle.travel for a number of years and am tremendously grateful for the amazing service it provides. I use it for everything from quieter ways to navigate my city to planning the route for bicycle touring.
I apologise if this has been asked before but I have not found it on the website
I have been thinking recently of how the algorithm chooses particular routes. Would you be able to provide some information on this? I understand some of this may be proprietary, in which case a high level description would suffice. I think it would be useful to understand the assumptions/limitations of the software to inform my own route mapping.
Thanks
Neil
Comments
Essentially every single segment (or ‘edge’ to use the technical term!) of road/path is scored. The score starts off as the pure distance from point to point, but is then adjusted for many many factors – surface (smooth is better), gradient (flatter is better), traffic levels, scenic appeal and so on.
When you ask for a route, cycle.travel then looks at all the possible permutations between A and B, and chooses the one with the best score.
The core map data comes from OpenStreetMap, the community-made map of the world. If the data is missing or faulty in OSM, that will show through in cycle.travel’s routing (e.g. if someone has mistakenly marked a stile on a cyclable path, cycle.travel won’t route through that point).
There’s a bit more about it in this talk that I gave to the Cycle Touring Festival the other month: https://www.youtube.com/watch?v=MB-Rvt5qBFo