Getting cycle.travel route planner to follow MY route choice

[mjr]

Is it possible to switch 'go direct' on for a portion of a route, then switch it off again once cycle.travel will plot ok? I seem only to be able to have it on or off for the entire route, thus meaning it just draws a straight line from start to finish, rather than over specific parts.

Yes: position via points either side of the problem, then click the lowest-numbered one and pick "go direct" from the pop-up.

[mikeymo]

What it seems to stem from is that the path/route in the underlying Openstreetmap (OSM) layer has not had a flag set as being suitable as a cycle route.

Which flag in OSM does that?

Ideally, bicycle=yes, permissive or designated, but the routing apps try to infer it from other tags if absent. https://cycle.travel/advice/map/details describes more.

Thanks. Your link describes how Cycle.travel interprets OSM tags. I was asking what flag (tag?) sets the section "as being suitable as a cycle route." within OSM.

"bicycle=yes" seems to mainly describe the legality of cycling on a particular section:

https://wiki.openstreetmap.org/wiki/Tag:bicycle=yes

as opposed to the physical suitability of a route.

I couldn't see a tag like that. Which makes sense, as what is "cycleable" will vary from cyclist to cyclist, so is to some extent a judgement call.

There's an OSM feature near me, which is a bridleway. I couldn't cycle on it. I edited it in OSM. Interestingly a couple of members of a local cycle group think it's cycleable (though on MTB/Gravel bikes). I edited the MTB tags in OSM to that suggested by one of these cyclists.

Cyclestreets still puts that section on a cycle route. Though cycle.travel and OSMand don't. It's a lesson for me, not to use Cycle Streets unless I can examine the route. I thought as an experiment to change the MTB tags to the most extreme, just to see if cyclestreets still sends "ordinary" cyclists that way.

[mikeymo]

If you are minded to do so, you can edit OSM and change/correct the flags on the path/s in question so that appropriate routing works. I think it takes a while to work it's way through the system, but is a possible long-term fix.

I believe the different routing apps and sites may still plot different routes, even working from the same OSM data.

In a way that makes sense. IF the tags in OSM are reliable (big if, obviously) then it would be perfectly possible for very different cycle route apps to be correct for their target audiences:

"Hardcore Mountain tough guys routing"

"Lycra Olympic wannabe on 22mm tyres and carbon thing"

"Tourist with 4 panniers, tent, kitchen sink"

Those three cyclists would probably want very different routes.

I don't know if the granularity, or even just plain accuracy, of OSM is good enough to allow market segmentation like that, but hopefully one day it will.

[Oldjohnw]

I am obviously being particularly thick here but if you want a route of your choice why use an app which has a set route?

[mjr]

I am obviously being particularly thick here but if you want a route of your choice why use an app which has a set route?

Lack of time to inspect every segment of every route.

[richardfm]

I am obviously being particularly thick here but if you want a route of your choice why use an app which has a set route?

I use cycle.travel to suggest a route and then add and drag via points to change the route if there is something I don't like. It's much quicker than starting from scratch and plotting an entire route myself.

[chris_suffolk]

I use cycle.travel to

a) suggest routes that I can then follow
b) create specific routes that I want to follow

Biggest benefit is abilty to easily download GPX file to Garmin, which is much better than the software that Garmin provides

[Mick F]

Use BikeHikeUK.
Simple and easy and foolproof, and has Ordnance Survey maps too.
http://www.bikehike.co.uk/mapview.php

[matt_twam_asi]

Which flag in OSM does that?

Ideally, bicycle=yes, permissive or designated, but the routing apps try to infer it from other tags if absent. https://cycle.travel/advice/map/details describes more.

Thanks. Your link describes how Cycle.travel interprets OSM tags. I was asking what flag (tag?) sets the section "as being suitable as a cycle route." within OSM.

"bicycle=yes" seems to mainly describe the legality of cycling on a particular section:

https://wiki.openstreetmap.org/wiki/Tag:bicycle=yes

as opposed to the physical suitability of a route.

I couldn't see a tag like that. Which makes sense, as what is "cycleable" will vary from cyclist to cyclist, so is to some extent a judgement call.

I think the closest are the tracktype, surface, and smoothness keys. I personally dislike tracktype and smoothness, and usually stick to surface as the most* objective. As you have suggested, subjective keys don't tend to get very far in OSM with good reason, they make assumptions about the data consumer that can't be known in advance.

*but not perfectly

There's an OSM feature near me, which is a bridleway. I couldn't cycle on it. I edited it in OSM. Interestingly a couple of members of a local cycle group think it's cycleable (though on MTB/Gravel bikes). I edited the MTB tags in OSM to that suggested by one of these cyclists.

Cyclestreets still puts that section on a cycle route. Though cycle.travel and OSMand don't. It's a lesson for me, not to use Cycle Streets unless I can examine the route. I thought as an experiment to change the MTB tags to the most extreme, just to see if cyclestreets still sends "ordinary" cyclists that way.

Please don't change tags just to test what a router will do. If a particular router makes an odd decision on a route, let them know.

[mikeymo]

Please don't change tags just to test what a router will do. If a particular router makes an odd decision on a route, let them know.

You might be right, that would be naughty of me, wouldn't it? But as it wouldn't affect the routers who do get it right, it wouldn't really matter, would it? I'd change it back later, promise

Or maybe cyclestreets could address the feedback I've left, twice in fact, just as you suggest.

This is actually an area I know quite well. Which may seem a bit weird, using a cycle route planner for somewhere I know. But at least it's given me a good idea of the different planners' reliability. I'd be cautious about using Cyclestreets if I didn't know the area.

[hemo]

Have you tried Kamoot ?:

[mjr]

Have you tried Kamoot ?:

Not recently. I don't like riding across ploughed fields!

[mikeymo]

Have you tried Kamoot ?:

Not recently. I don't like riding across ploughed fields!

I tried using it for my test not-suitable-for-ordinary-bike segment. Komoot sent my along it when I had preferences set to "touring". If I changed my preference to "road bike" Komoot avoided it. So there is some granularity there, more than the usual "paved only" option. I don't personally agree that the segment I chose is suitable for a touring cyclist, but at least there's the choice. Komoot has 5 different cycling preferences, Road, Touring, Mountain, Gravel and Enduro Mountain. So that's good. There's the option to view surface types, but as far as I can tell not to plan a route avoiding them.

I guess I'd have to explore more segments to see how Komoot handles them. But as most of the routers seem to be drawing from the same dataset I suppose if, for instance, Komoot thinks that an S1 surface of 7% gradient is OK for touring, then it will treat every other segment the same.

Komoot seems to have a sort of hybrid pricing. Free(for one small region)/Single Payment/Premium subscription. The premium pricing didn't look outrageously expensive, though I personally don't need it. I don't mind subscription pricing, and pay for some software I use on subscription.

[mikeymo]

RideWithGPS have announced a "popular routes" type thing:

https://ridewithgps.com/news/4859-introducing-popularity-influenced-routin?otu=

[Richard Fairhurst]

RideWithGPS have announced a "popular routes" type thing:

https://ridewithgps.com/news/4859-introducing-popularity-influenced-routin?otu=

Yep - they've been working on this for a while. My personal opinion (we discussed this a bit in the Bike Show podcast the other week) is that a good route is defined by the absence of cars rather than the presence of other bikes, and that popularity-based routing is really troublesome because it can trivially be broken by club runs and things like that - Strava suffers from this a lot. But fortunately I don't have a monopoly on opinions and it's good that there are other approaches to routing.

I'm playing around with some really interesting stuff for cycle.travel's routing which I don't believe anyone else is working on, but it won't be finished this summer because Mrs F insists that I have to go on holiday...