Hello. Didi here from Bike Citizens.
Thanks for your feedback. After looking at the cases mentioned in this thread, I'll try to give some answers:
Yes, we do use OpenStreetMap (as most cycle route planners).
We built a routing engine from scratch - mainly because there was no real alternative for offline routing when we started with this project.
@chdot
> Doesn't seem to have a 'start where you are' option.
The App uses the current location as starting point by default. Only if it's not available (e.g. no GPS fix) manual entry is needed in order to calculate a route
> Suspect not taking into account traffic lights and/or hills enough(?)
We do consider elevation data. We do not yet directly consider traffic lights, but look at the sizes of crossing ways in junctions (on average crossing a large road takes longer then crossing a small one).
@wingpig
> Presumably it edits out the bits couriers use which they should legally not use?
Yes, it does :-)
Of course courier style routes aren't a viable option for most of us. Still, they can give very useful input when it comes to fine tune the "cost factors" of the routing. They know very well what's less or more safe, where time is lost etc. Also, two of us regularly ride with a trailer with Baby/kid inside, so we are also aware of different perspectives / requirements.
@tk
> I presume they are trying to do this to detect which cycle routes are used but from what I've heard most people are turning it on in a car cradle while driving to get the free maps quickly.
We do use the data for visualizations (e.g. http://www.naden.de/blog/bbvideo-bbpress-video-plugin -->
However the "cylce to free" option was not created with that in mind. It was our solution to the dilemma that the previous classical trial period (limited to 30 km) was a disincentive to cycling.
If people use the car to unlock the maps, that's a pity. We may later retrospectively analyse the tracks and get back to those who cheated ;-)
@chrisfl
> Although they do allow you to choose a bike type MTB, City Bike or Road Bike.
We do. It's just a bit hidden under the "More" button. However the bike type selection doesn't so far have a big influence on the routing - except when it's about the surface quality (and the correct data is in the OpenStreetMap).
> If I look at Waterloo Place to Haymarket, Cyclestreets says 12 minutes (9 traffic lights and 4 pedestrian crossings) while bikecitizen says 7 minutes.
It says 7 minutes with the "fast" routing profile
With the "balanced" one it says 10 and with the "easy" one 12 minutes. We assume different average speeds based on routing profile selection.
> Also looks like hills aren't taken into account, so bottom of the mound down to the start of lower Granton Road, bikecitizens gives 12 minutes for each direction.
I can't confirm this. See downhill (13 min), uphill (18 min) - both with "balanced" profile (with the "fast" one it's 9 and 12 mins). Or did I misinterpret the start/destination locations?
> Harrison Road to Murieston Road seems to refuse to go via Ardmillan Terrace.
According to the map data this decision looks reasonable to me. Ardmillan Terrace is tagged as "Primary" street in OpenStreetMap while the alternative route consists of "Tertiary" and "Residential" streets. Our router tries hard to avoid Primary streets in all but the "Fast" routing profile.
I can't however tell for sure if it's a good decision in this concrete case (you seem not to think so). It may be that in Edinbourgh (or probably the UK in general) primaries aren't as "bad" (from cyclists perspective) as they use to be in Austria and Germany. If that's the case, we will take a closer look and fine tune the parameters for this region.
Do you think the detour is reasonable at least with the "easy" profile (think about riding with kids for example)?
posts
