Route GuidesRoutes City GuidesCities Map My

Climb inconsistencies

Tuesday 14 March
by Miguel
in forum cycle.travel
Find a better bike route. Try our map & route-planner »

Hi,

Having one GPX file that I uploaded to different sites, I found that none results on the same meters climbing and sometimes also different distance.

Can you please explain what system does cycle.travel uses to calculate climbs and why the results differ so much?

For instance the GPX of: http://www.gpsies.com/map.do?fileId=ffylhuwcnnwyuqdm

does result on:

Google Maps: 75.7 km / 1378m up / 1487m down

GPSies: 75.75 km / 1283m up / 1390m down

Bikemap: 75.7 km / 1150m up / 1250m down

Cycle.travel: 82 km / 1500 m up

Cheers,

 Miguel Anjo

Comments

Wed 15 Mar, 08:53

Hi Miguel,

It's not so much that the climb calculation is different - it’s more about the uploaded route itself.

When you upload a GPX with cycle.travel, it actually tries to reconstruct the via points for the route you cycled. That means the route is editable, you get turn-by-turn directions, and so on.

It doesn’t (yet!) always get these right so you’ll sometimes see little side-trips off the main route where the via point has been placed wrongly. So, for example, at http://cycle.travel/map/journey/35541, you'll see that via point 9 has been placed slightly off the route, causing a detour up a side-road; and the route from point 6 to point 3 is wrong, because cycle.travel wouldn’t usually choose a busy road like the N-634 so it’s got a bit confused. (It looks like there’s a painted bike lane there – is that right? If so then I can mark it in OpenStreetMap and cycle.travel will be happier to take it.)

Once the route is corrected then the climb/distance will be pretty much the same.

I’m continuing to work on getting the via points as close to the original route as possible but it’s a hard problem! :)

Thu 16 Mar, 18:56

Hi Richard,

Thanks a lot for your explanantion. I can see now how cycle.travel works a bit better. Klaus from GPSies.com also sent me a link with a lot of explanations how they calculate the altitude data for GPS tracks (in German):

http://blog.gpsies.com/article/206/hoehenmeter-gpsies-sucht-die-wahrheit

The GPX I sent you was created on GPSies and had already some problems as you spotted. The point 9 was indeed outside a road and while GPSies just makes a "U-turn" in place, cycle.travel tries to get to the closest place to make a legal U-turn, thus increasing distance and climb.

Then I was thinking cycle.travel would believe that all the GPS points on a track represent "turns" and would just connect them with a straight line not looking if there is a road or not.

For the specific case, the route from point 6 to 3, I make specifically to go via the N-634. This because by experience in Spain the national roads next to highways have always a big shoulder and are mostly traffic free, making a good road to cycle instead of going zig-zag around small roads.

Because of this, on our trips in Europe these two last years, we would use cycle.travel for France and Portugal, but in Spain it does not give good results. Small Spanish roads tend to have very steep hills and national roads are usually very good and safe for cycling. But I understand that it is impossible for you to know this country specific details.

Cheers,

Miguel Anjo

Thu 16 Mar, 21:12

Funny that you mention that! As of the latest update (a week ago) cycle.travel now uses real traffic data for many Spanish roads - just as it does in the UK and US. So it will now often choose an N- road where it wouldn't have chosen one previously. For example, if you ask for a route from Santander to Burgos, it will follow the N-623 for much of the way, because the number of cars is very low.

The N-634 has higher traffic so cycle.travel is less keen to follow it. But adding a shoulder/bike lane tag in OpenStreetMap will make it more likely.