Is there any agreed or proposed practice for tagging cycling/walking infrastructure that connects population centers? I’m thinking in terms of what would a cycling-oriented map want to render at around zoom 9, much before all the cyling infrastructure inside towns gets dumped on the map.
I’m specifically not talking about signposted routes or networks, but rather any built cycling infrastructure (highway=cycleway or similar) outside towns and villages. In my area these typically follow tertiary roads, in other places these could be bike paths along canals etc.
I think we have something of a communications failure here ! Returning to the question:
Zoom 9 is “zoomed quite a long way out”. In that example, looking between Helsinki / Hämeenlinna route, at zoom 9 OSM’s “Standard” layer just shows the motorway. However, if you were cycling it, you might instead choose the secondary road with features useful to cyclists. I’d suggest that you ask yourself what features are useful to you as a long-distance cyclist**, ask yourself how they are tagged in OSM, and then make sure that your map shows those features - and makes sure that it is more visible at zoom 9 than the adjacent motorway.
Neither of osm.org’s two cycling maps does this at this zoom level - as you’ve noted, they show routes rather than infrastructure information.
** and that might vary by region - “winter service” is probably going to be a higher priority on that road than some other places.
Using Helsinki to Hämeenlinna as an example, if I were a mapping software I would want to show the cycleways that run all the way from Helsinki to Kerava to Järvenpää to Jokela to Hyvinkää to Riihimäki to Hausjärven kirkonkylä, and then from Turenki to Hämeenlinna.
The main problem I’m getting at here, is that even where this type of inrastructure exists, it is not easily discoverable, even on dedicated cycling maps. When CyclOSM finally starts showing cycleways at zoom 11, it’s as really thin lines so that towns wouldn’t become a complete mess.
Thank you for the link, I was not familiar with Richard or his map.
It appears this map renders all cycleways at once at zoom 12. At zoom 8, where you linked to, all I can see is some Eurovelo routes (that do not necessarily correspond to any actual cycling infrastructure).
Basically, this doesn’t demonstrate the thing I’m asking about, at least not in this area.
Now, you are all right to question the approach I asked about. I do not particularly want to go about adding extra tags, and I would welcome coding magic handling everything instead. I also have no demands to Richard about what his map specifically should render.
But I’m proposing that
“show cycleway connections between towns at low zooms” is a useful feature for cycling maps
there doesn’t appear to be tagging to facilitate this, as there is for roads
I have not seen a successful demonstration of this using any alternative method to tags
if such a method is possible, for some reason even commercial solutions targeting cyclists do not implement it, suggesting there is a downside to it
Sounds like what you are asking for are either cycle routes or some heuristic to automatically detect cycleway connections between towns. It seems you also only want to find those routes that have dedicated cycle infrastructure?
You could of course filter the existing routes by ‘bicycle friendliness’ and only pick those that use dedicated infrastructure for most of the way.
Defining and finding bicycle infrastructure is not particularly easy though. A minor road with little car traffic can be just as bicycle friendly as ‘dedicated’ cycleways.
there doesn’t appear to be tagging to facilitate this, as there is for roads
Roads are always car infrastructure, so a router/map for cars just doesn’t face the same challenges as for bicycles. The available tagging schemes are the same for cars and bicycles.
I agree this is a challenge. I proposed a hierarchy for cycleways the same way there is for carriageways, but this was mostly rejected by the community.
I don’t see a good way to solve it within today’s framework without a lot of meta analysis.
Maybe I’m being unclear and for that I’m sorry. I’m specifically concerned with maps and rendering, not routing software.
I understand that some countries have official routes and node networks to solve the core issue (discoverable, safe routes for bicycle traffic). I’m vaguely familiar with how those are presented in OSM.
I happen to live in an area which doesn’t have that, but does have actual highway=cycleway tagged cycleways between towns. So I’m trying to understand how I can work with that, to make these more discoverable, as right now they’re rather hard to find. Even on dedicated cycling maps, as demonstrated above.
I agree that comprehensive cycle routing would ideally include roads designated as ”safe”, but I don’t have any source of authority to refer to here. But cycleways objectively exist.
What you suggest might be difficult sounds exactly like what I do for the maps I create for pedestrians. If you have a look at the map legend here and look up and down, you can see “different classes” based only on OSM tags - in this case obscure England-and-Wales-specific legal access rights, but it could easily be computed from “a bunch of tags indicating how easy it is to cycle somewhere”. If you move left and right you can see how other attributes are handled. Some are just one OSM tag (e.g. “is this in a tunnel”) but some are calculated from many, such as the “less visible” paths indicator.
You’re absolutely correct that existing cycling maps don’t make the infrastructure that you are interested in easy to find at low zoom levels. However, I suspect that it’s easier than you think it is to create a map showing it - what I do for pedestrian maps works for both raster maps like the ones you see on osm.org and vector maps too (it’s even mostly the same code), and I’m sure that you could do something similar.
The cycleways I’m looking at just don’t have all that many tags. Cycleway with moped=yes would be a clue (locally, but not necessarily in other countries), and I guess so are relations to the few routes (state=disused) we have. But there are sections that are literally just
which is about as unremarkable as it gets. If rendering has access to logic like “is this cycleway near a main/secondary/tertiary road with speed limit >= 60” then that would (locally) be a fairly good clue.
But if comes down to stuff like this, wouldn’t a dedicated tag be better than having mapping software create “best guess” local rules for every area they try to support?
Obviously I was just lucky when I picked a random road in Finland!
It’s possible (but far less straightforward than what I described earlier) to do that sort of thing. I’m unsure whether you would consider that a “good” or “bad” place to cycle, by the way, but that’s not really relevant…
Even just that might be enough to get you started.
Just chiming in as a fellow Finn and an avid cyclist who often cycles also longer stretches between cities.
Many people have made good suggestions already, so I’m not sure how helpful I can be.
However—just as a hint—the Finnish OSM community has been in the process of reworking instructions and examples on how to tag cycling and walking infrastructure (kevyt liikenne in Finnish) in Finland. This work started partly because on the first of June 2020 the new and vast traffic law reform (Tieliikennelaki) was enacted and there were e.g. a bunch of new traffic signs that came into existence. Partly this exercise exists because tagging cycleways and footpaths in Finland was and is quite diverse and variegated (to put it politely).
You might want to take a look at the (very lengthy) discussion, or directly at the still incomplete proposal for the new instructions on the Wiki (in Finnish). There are lots of details to hash out still (and the path controversy features prominently there as well). But we are working towards a reasonable compromise.
Most cycleways here are paved, and some that run in the countryside are not. I think whether or not there’s a surface tag is mostly about who tagged it and when, or at least I can’t see any pattern to it. (Unrelated to this topic but the main reason I participate in OSM is to fix missing/incorrect surface tags, as certain global apps make locally incorrect assumptions about surface, causing problems for road bike routing.)
Also to note, I’m not trying to argue anything about ”good riding”. I think for cycleways ”safe riding” is a reasonably objective characterisation, and that is what some people are looking for. Commuters, families, new and casual hobbyists…
Indeed, you might already find a few comments from me in that thread
My concern here does not have any basis on local law or traffic signs, so I don’t see it just as a local problem. E.g. from traffic signs and external features perspective, the cycleway along Eriksnäsintie in Järvenpää is just a normal cycleway. The only thing that makes it special is that it’s the only way to get from Järvenpää to Jokela and eventually Hyvinkää on a cycleway.
I feel knowing this kind of thing about a cycleway has value to users of OSM, but it’s a slightly different topic than trying to map a local traffic law (and its imperfect execution) to OSM metadata. Still, I’m happy to discuss this in the Finnish thread too if you feel it fits there. A local solution is better than nothing, especially now that we no longer have to rely on global mapping apps.
Here’s my cycle map, and I chose this zoom level because this is where I render cycleways that I classify as “trunk” or “express” based on tagging.
There are several cycleways that should have been rendered at this zoom level because they are regional links (within the bounding box of the map).
Based on segment tagging, they are indistinguishable from cycleways of absolutely no regional relevance that I do not want to render at this zoom level.
Looking at the tags of individual segments is an assessment of quality, and just like carriageways there is a non-linear relationship between quality and function.
Some trunk carriageways in Norway are narrower than my inner city residential street, but because they are tagged as trunk, they are rendered at low zoom levels.
I could elevate cycleways that are part of a route relation. This has been considered by the OP as not applicable to him. For me it is, but it’s a huge task, especially compared to what I would face if I were rendering a car map. I can’t imagine that this would be accepted if it were the same for carriageways…
As I mentioned above, you can do this by running a query against built-up area polygons.
I do that on cycle.travel for other purposes. If you look at (say) Bike map | Cycle route planner | cycle.travel you will see that cafes and pubs in small villages and rural areas are shown, whereas those in towns and cities are not. You have to zoom in further to see those in towns and cities. Similarly, if you look carefully you’ll see service roads get similar treatment, there are subtle styling differences between roads in urban and rural areas, and so on.
I don’t do it for cycle infrastructure below z12 because in most countries, I believe the national/regional cycle networks (as mapped with route relations) serve this purpose better. But that’s just a personal view and you could absolutely use this technique to show traffic-free cycleways in rural areas at lower zoom levels if you wanted to. There is no need to invent new tagging for this, especially for something that isn’t an on-the-ground fact.
(What I would really like to do is derive the de facto “trunk” routes for cycling from a routing graph, and then highlight those on a map. I’m pretty sure it’s possible, I just haven’t worked out a performant way of doing it yet.)
Thank you for explaining what you’re doing. For the record I think your map looks absolutely brilliant and the idea of showing cafes in small villages is excellent. (I have a much more inelegant personal solution of maintaining a cafe list in Google’s My Maps.)
Since I do not have a map of my own, despite people suggesting otherwise , I cannot really evaluate how feasible your solution is for others to copy. As in, is everyone else just overlooking an obvious method or is this something particularly clever or difficult.
From a very brief look, it seems that there are areas where the logic doesn’t work as expected (small patches of major cities, or indeed the entire town of Inca, Mallorca). Am I wrong to assume you’ve spent quite a bit of effort fine-tuning this for the areas your map mainly focuses on?
Anyway, I can agree that for something like showing village cafes earlier, it’s not really reasonable to expect tag support. But I strongly feel that ”trunk cycleways” as a concept should be so easy to distinguish that even mainstream maps could support it, without local tweaks.
Is there some reason this is not desirable that I’m failing to understand?
! Returning to the question:
