Talk:Key:traffic

Verifiability

How mapper is supposed to verify or survey this? It seems completely mismatching OpenStreetMap Mateusz Konieczny (talk) 20:25, 16 January 2021 (UTC)

Hello ! In France, local and national authorities are doing regular surveys about this ;) Djakk (talk) 10:18, 17 January 2021 (UTC)

So it would be importing of data unverifiable by mappers? Mateusz Konieczny (talk) 10:59, 17 January 2021 (UTC)

Well, mappers can check the data (then you have to trust officials like for some unsigned landmarks : https://www.openstreetmap.org/node/4896032585) or feel the amount of traffic by driving on the selected road. It is as subjective as the difference between highway="secondary" or highway="primary" in France. Djakk (talk) 13:20, 17 January 2021 (UTC)

Names can be verified by asking people living around given location and it is feasible for mapper to do this. Making systematic measurement of traffic is not doable Mateusz Konieczny (talk) 13:27, 17 January 2021 (UTC)

Mappers can also manually count the number of vehicles passing by (during 10 minutes, then *6 *12 to get the average amount of traffic between 7am and 7pm). Djakk (talk) 09:44, 18 January 2021 (UTC)

That is ridiculously unrepresentative - there are typically massive changes in traffic during day, sometimes also between days Mateusz Konieczny (talk) 10:00, 18 January 2021 (UTC)

That is true, but with this method I think you can still get an order of magnitude. Djakk (talk) 11:13, 19 January 2021 (UTC)

"order of magnitude" - that should doable just from highway=* and from location Mateusz Konieczny (talk) 12:37, 19 January 2021 (UTC)

I also support this tag (against Mateusz' opinion; he seems to be against all kinds of traffic related tagging which is not directly verifiable [which is by my opinion is unverifiable on the place just like any administration boundary for example]). This measurement is published by several organisations and could be even automagically updated, provided we tag in some kind of stable reference. This information would help tremendously bike routing, for example, in case of no other meaningful tags (like highway=unclassified which is fitting - but not designated - for bike traffic). --grin ✎ 14:43, 5 September 2022 (UTC)

"just as any administration boundary for example" - this is not my position Mateusz Konieczny (talk) 14:55, 5 September 2022 (UTC)

I have updated the bad phrasing. Thanks. --grin ✎ 15:04, 5 September 2022 (UTC)

"any administration boundary" is going too far. For start national borders often are well signed, lower level are also sometimes fixed and/or verifiable by local survey doable by regular humans. For very low levels and unmarked national boundaries I am fine with super-special exemption for them, but I am not fine with such super-special exemption for this kind of data that cannot be ever improved by normal people (unlike official boundary data) Mateusz Konieczny (talk) 16:54, 5 September 2022 (UTC)

Trunk?

I've never heard 'trunk' being used as a quantitative adjective before. Is there any evidence of it being used to mean 'very heavy' outside of OSM?

This traffic tags appears very subjective. In what form is the French data supplied? Is it using adjectives or cardinal numbers? Is it over the whole length of road or split into sections? Does it account for the time of day? the highway=* tag gives as much idea about the amount of traffic as 'traffic' does.--DaveF63 (talk) 09:37, 7 February 2021 (UTC)

Ambiguous value

Amount of car on a way in some period does not actually represent traffic jams on it. Imagine a road with traffic or traffic:hourly=10. Does it mean that there were a huge jam on it during this hour, or it mean the road is underused, so only 10 cars passed through it?
Value which is really useful is the average speed of vehicles during period, although actual traffic value also may be useful.
This tag was most likely supposed to be calculated automatically or obtained from government sources, so it does not seem that speed measurement will become something more difficult.
--Richard Try (talk) 10:31, 18 February 2023 (UTC)