Is there an online route planning website that takes into account local knowledg

[Russell]

e?

That finishes off the subject line!

Yesterday we rode up out of Brighton to ride the south face of Ditchling Beacon.  Ditchling was a piece of cake compared to the climbs (and descents) to get there.  I plotted a route on Viewranger which on paper looked OK.  Ended up on a (for us) near the limit climb on a busy road in 28deg thinking there has to be a better route.  This followed several other near the limit climbs and steep descents.

Is there a way of getting local knowledge into the route planning stage to avoid what look like good routes but turn out not to be.

Thanks

[robgul]

. . . I don't want to appear rude here - BUT

Using the Ordnace Survey maps and looking at the brown contour lines (proximity to each other, their numbers in metres above sea level) plus the black dots with "spot heights" which are pretty plentiful, with a little bit of experience, will give you your own "local knowledge".    I'm not familiar with ViewRanger but most mapping systems will allow you to look at a strip route profile which, while not ideal, does give you a fair indication of the ups and downs - and how big they are.

As always, like beauty, severity of hills is in the eye of the beholder!

[grams]

What kind of local knowledge? Strava has a “popularity” option but that’s based on roads popular with Strava users, which skew a certain way.

There are other opinionated route planners, like cycle.travel which strongly favour low traffic roads, sometimes to the point of Sustrans silliness (though it at least has a “paved only” option).

So as always, what makes a good road for cycling will be completely different depending on who you ask, so it’s not clear what specifically you’re asking for.

[MikeFromLFE]

I've often mused on this very topic, and concluded that it's an impossible thing however desirable.
If you look for popular routes you end up in the Strava dominated crowd, which is great if you're looking for a fast route and / or segments.
If you look for quiet routes (cycle.travel) you might increase your distance by a margin you are unhappy with.
Flat routes / fast routes can lead you onto either extreme of canal mudpaths or psuedomotorways.
You might like cycle paths, I may abhor them.
You might want scenery, I might want urban grandeur.
There's no pleasing cyclists, because there's no 'average' cyclist, and long may that continue.

[freeflow]

It is a very simple endeavour to generate your own local knowledge.  It just takes a small amount of effort.  For my purposes I've found that the free version of Ride with Gps is perfectly acceptable.


•  Set the map to cycling mode and follow roads. Plot your route
•  Determine the nature (size) of roads by switching maps (the OSM layer is good for this)
•  Drag the street view man over the map.  Roads along which the Google camera has been driven turn blue.  If your route goes where the roads are not blue you need to check in more detail
•  Drop the man onto the map to check the specific nature of the road/junctions.  I do this a lot for sections of bigger roads to check if there is a cycle path
•  If you set on using a route which does not have blue roads or are othwerwise concerned, publish a link to the route and ask for **specific** comments/alternatives.  I've done this for a few long distance routes I've plotted.

[robgul]

It is a very simple endeavour to generate your own local knowledge.  It just takes a small amount of effort.  For my purposes I've found that the free version of Ride with Gps is perfectly acceptable.


•  Set the map to cycling mode and follow roads. Plot your route
•  Determine the nature (size) of roads by switching maps (the OSM layer is good for this)
•  Drag the street view man over the map.  Roads along which the Google camera has been driven turn blue.  If your route goes where the roads are not blue you need to check in more detail
•  Drop the man onto the map to check the specific nature of the road/junctions.  I do this a lot for sections of bigger roads to check if there is a cycle path
•  If you set on using a route which does not have blue roads or are othwerwise concerned, publish a link to the route and ask for **specific** comments/alternatives.  I've done this for a few long distance routes I've plotted.

Agreed on all that and I use that plus other plotting systems - unfortunately none of the on-screen "look at the road" options show what the hills really look like  - Streetview is brilliant (although it doesn't work in Germany as I found out the other day) - but other than reading the contours and to some extent the profile options I can't see an answer.

[drossall]

It is a very simple endeavour to generate your own local knowledge.

It's quite hard if you're not talking about your local area. One of the brilliant things about Audax is getting the benefit of other people's local knowledge in areas that you do not know well. And even minor roads can be rat runs that only a local would know to avoid.

The question sounded pretty reasonable to me. Doesn't CycleStreets use input from real riders, for example?

[robgul]

It is a very simple endeavour to generate your own local knowledge.

It's quite hard if you're not talking about your local area. One of the brilliant things about Audax is getting the benefit of other people's local knowledge in areas that you do not know well. And even minor roads can be rat runs that only a local would know to avoid.

The question sounded pretty reasonable to me. Doesn't CycleStreets use input from real riders, for example?

I'm not too sure about CycleStreets - each time I've used it the routes are pretty much the same as Google Maps directions when set for a bike - lots of roads you wouldn't want to ride on in your right mind!   I've only ever used to get a global picture and then refine my planned route using map-reading skills and local knowledge.

A prime example in my vicinity is the invariable recommendation to cycle on the Fosseway in Gloucestershire, Warwickshire and Leicestershire - the road is littered with "n accidents here in .... " signs etc.  To be avoided at all costs on a bike.

[FifeingEejit]

Doesn't CycleStreets use input from real riders, for example?

In theory so too does OSM and there's a load of crap in that too.

[Peat]

I believe this is what KOMOOT is trying to achieve. Their routing algorithm is influenced by their users logged rides.

[L CC]

I believe this is what KOMOOT is trying to achieve. Their routing algorithm is influenced by their users logged rides.

So is Strava's - but I don't always want to ride someone else's commute or hill repeats route.

Best is really best fit-for-you. No single algorithm can manage that.

[Lightning Phil]

Generating the edge data behind the routing is expensive, so that the route queries are cheap.  Generating the data takes a few hours (certainly from Osm raw data) unless you throw lots of computing grunt at it.  The edge data is a few GB for the UK. The routing queries are sub second.

You are not going to get personalised routing. Every option you see on the online routing involves pre generating a fresh set of edge data and weightings etc  You can’t do that on the fly (well maybe a dedicated supercomputer can).

The best you can hope is that a routing engine comes up with something close to your preferences and that you can drag the routing bits you don’t like around on a map.

[psyclist]

You are not going to get personalised routing.

To an extent, Kamoot gives you a personalised route based on your preferred type of riding (bike touring, mountain biking, road cycling, gravel riding, enduro mountain biking) plus whether you are riding an e-bike. Extensions to this model could include additional choices, such as hill avoidance, avoidance of roads with accident blackspots, etc, making for a route that would be quite bespoke for an individuals preferences.