Hello everyone,
I recently discovered cycle.travel, and I'm now testing it. I tested the round-trip suggestion feature in France, and if it's working way better than other apps, I got a bit surprised by 2 routing decision cycle.travel took.
The first surprising routing is reproduced in this example journey: https://cycle.travel/map/journey/648597 - do you understand why the app route people through the D 448 to Boulevard Charles de Gaule (i.e. the second part of the routing, not the first part) instead of passing by the bicycles path number 3 (also named "Scandibérique") ? The D 448 is terrible for bicycles: lots of car traffic, fast cars, stressing placing on the road with all the accesses to the "La Francillienne" major road, while the "Scandibérique" is clearly mapped on Open Street Map (I checked), and marked as accessible for bicycles.
The second surprising routing is reproduced in this second example journey: https://cycle.travel/map/journey/648610 - do you understand why it makes this back-and-forth in the "Allée du Bois des Folies" ? (we don't see the name of the street in cycle.travel map, but it's this little perpendicular dead-end street we can see along the way).
Thank you in advance !
Comments
Both of these are (I think) issues with OpenStreetMap data.
In the first example, the D 448 is marked in OSM as being part of Veloroute V33. cycle.travel assumes that, in France at least, a road that’s been designated as one of the national veloroutes is probably ok, so it’s happy to route along there.
My suspicion is that OSM is wrong/outdated here, and that V33 actually follows the riverside path along with Eurovelo 3. Correcting this in OSM should fix the issue in c.t next time it takes an update.
In the second example, OSM weirdly has a turn restriction marked at this junction, saying that it’s forbidden for any traffic to go straight on at the junction. Consequently c.t turns left, does a U-turn, then turns right to avoid this restriction! I’m sure this must be wrong, and it seems odd given that there’s an gate preventing cars a few metres further on anyway.
I’m happy to fix these in OSM (or you can if you’re comfortable with editing OSM), but if you could confirm my assumptions about what’s actually there are correct, that’d be super helpful.
Hi Richard, thank you for your answer.
> In the first example, the D 448 is marked in OSM as being part of Veloroute V33. cycle.travel assumes that, in France at least, a road that’s been designated as one of the national veloroutes is probably ok, so it’s happy to route along there.
> My suspicion is that OSM is wrong/outdated here, and that V33 actually follows the riverside path along with Eurovelo 3.
It makes sense, I will check this point, it's the first time I pass by here, so I don't know. The only thing that I can tell is if D 448 is indeed open for bicycles, it's a terrible section ^^'
> In the second example, OSM weirdly has a turn restriction marked at this junction, saying that it’s forbidden for any traffic to go straight on at the junction. Consequently c.t turns left, does a U-turn, then turns right to avoid this restriction! I’m sure this must be wrong, and it seems odd given that there’s an gate preventing cars a few metres further on anyway.
Ah yes, I see it too in OSM now, and yes it's clearly wrong, I passed by here yesterday. I will fix, I'm confortable with that.
Thanks again for your help !
Okay, you are right for the V33, it should pass along the Eurovelo 3 on this section (https://www.af3v.org/les-voies-vertes/voies/929-v33-la-seine-a-velo-de-vernon-aux-andelys/#headCart -> Corbeil-Essonnes), and not the D 448. I will make the fix.
I fixed the second issue.
Thanks.
I fixed V33 / Eurovelo 3 as well. I will check the routing are fixed once updated on cycle.travel.
Superb - thank you! Next mapping update will be in a couple of weeks.
For the record, I checked: the 2 issues are now properly fixed.