Closed FFMbyBicycle closed 1 year ago
The estimated_traffic_class is generated by the estimated_traffic_class function and underlying code.
It is not clear to me how this code is working but I think that with enough time you can find out what it is doing. Documenting that would be great.
@polyscias, Arndt once described it here (though in German): https://forum.openstreetmap.org/viewtopic.php?id=31044
The estimations are working fine for me
Yes, but you have to be careful when using it, because although it works quite well, there is still a considerable percentage of busy roads that are not classified correctly.
Okay, I did put the text through deepl and the idea is clear to me.
I am not so sure it is a good way to get the "estimated traffic load" is there within a city more traffic on the roads getting to the highway entrance/exit? I think it is better to use the OSM highway qualification (primary/secondary/tertiary/unclassified/residential) and maybe the maxspeed to get a "estimated traffic load"
In Dutch context, a busy road (with >=50 km/h speed limits) is often chosen for bicycle routing because there will be cycle paths alone that road.
Thanks for your hints! The system is not perfect, but working quite good in my opinion.
I didn't want to start a discussion about other implementations, but get it working for bikerouter / other instances. Currently you get different results using the same profile depending on which brouter-web instance you're using and that's a little bit confusing. Sadly I'm not a software engineer and can't reverse-engineer the final segment data provided by Arndt.
This works for me but I have no idea what we have to do with the generated .trf files.
Looking at the code, I would say that these .trf files are ingested automatically by the addTrafficClass
function:
Hello,
Thank for the documentation above! I wasa interested to discover the jobs and could generate the RD5's with the "estimated_traffic_class" tag by following the steps decribed...but I think, there is an error somewhere...
Assuming the generation is done in the "root" directory mapcreation, then we have these commands: cd tmp mkdir traffic then start the seek jobs to generate the trf files in mapcreation/tmp/traffic
Anf then start the "waylinker" to add the tags: But on my local test system, waylinker is looking for trf files in mapcreation/traffic, NOT in mappcresation/tmp/traffic !
After moving the traffic directory to mapcreation, the tags were added properly
The estimated_traffic_class
data is now generated as part of the new pseudo tag generation.
Preliminary documentation: https://github.com/abrensch/brouter/blob/master/misc/scripts/mapcreation/readme_database.md
Together with @mjaschen I tried to find a way to generate the "estimated_traffic_class" data from scratch. The estimations are working fine for me and I'm missing them at bikerouter.de :wink: At the moment I have to decide if I use the traffic data at brouter.de or if I will be able to send a Link to someone with my profile.
So far we found out that there is a car-traffic_analysis.brf which seems to be used for this and calls a special case using "seed" keyword. This generates some data at a "traffic" subdirectory. This works for me but I have no idea what we have to do with the generated .trf files.
There seems to be a traffic_simulation.sh outside of this repository, maybe this does the magic? Or am I missing a step somewhere?
I can offer some Linux/scripting knowledge, time and will to write some documentation.