We have a cabin that we can usually get two bars of Verizon coverage if we are outside the building. But for a direct line of sight to the only mobile tower in the area I would need a 180 ft (55 m) tall antenna. All of our coverage is apparently due to signal being reflected off one of the surrounding mountains. And that signal varies over time, we went a month last winter when there was no reception at all.
That one mobile tower is the only one in the area and only serves Verizon so if you have a different network you will no coverage at all.
Seems like JFDI is what the OP said is intended. Perhaps weâve dissuaded what seems like a hopeless and likely doomed-from-the-get-go scheme. Especially that âthis is already mappedâ (by the carriers, junkily with noise, which is the best THEY can do, together with their lawyers, because theyâve thought about how to map these data, so they do a âmehâ job of it). Perhaps not and we get JFDI and that dead-ends, as it seems destined to do.
We (other posters here) are your friends, who are leading you âto waterâ (as if a horse, as the saying goes), AWAY from doing this, as it seems a foolâs errand. Not your adversaries (though nobody likes to see detritus leftover eraser crumbs that have not been blown away). Our map is plastic enough to absorb such data, which it does from time to time. Though, in this case, because you asked (via Proposal), people here seem to think this folly. You could listen to us, you could experiment towards a dead-end, up to you. We seem to have told you what we think. (Though more perspectives are welcome).
Whatever your level of skepticism toward this initiative, I think itâs sufficient to lay out your concerns without rubbing it in. If youâre concerned about this kind of mapping causing harm, that would be a different story, but so far the feedback only seems to be about how this kind of mapping is likely to remain in the realm of exotic micromapping, which is not a problem in and of itself.
I said it, without meaning for it to sting. If it stings because it is true, I did not sugar-coat it. If offense was taken (none meant), I will apologize, but Iâll hold off on offering an apology unless and until if I am asked for one.
If somebody brings an idea that others say âthis is a bad ideaâ or âwe donât do thisâ or âwe do this, but weâve also seen things LIKE this go south quite fastâŠâ weâre in a world of adults. No unkindness was meant.
I donât see any harm in this kind of mapping, and as a data user, I might find it helpful. I understand that there are a lot of things we could do to be more specific, e.g. which carriers offer service at the campground, but a general yes/no as to whether one is likely to be able to make a call from a given campground is of use.
I share the concerns expressed, and suspect the data would be generally of limited value. One thing that might help increase the value is, when adding it, only do so when itâs unexpected. I.e. a campsite in the middle of nowhere that surprisingly has coverage, or a building that unexpectedly acts as a Faraday cage. That could be at least of some interest to people (albeit the value would also depend on good changeset comments explaining how the tests were done).
I appreciate the sentiment! I never thought I would be some off put from OSM by proposing some change. Even the Vice-President of the US board emailed back with simply: âThis isnât really appropriate data for OSM, sorryâ and hasnât given any reasoning or feedback.
I guess JFDI is the way I most go because thereâs a significant lack of intent to create a beneficial proposal. I understand concerns expressed with verifiability and variability in data. For now I will be editing the proposal with the following key items. If anyone has beneficial feedback to improve the system it would be greatly appreciated. Otherwise the community can reject the proposal and Iâll JFDI this way.
Have the initial four tag values for cell_reception âyes,assumed,limited,noâ
Have only the following additional keys available: data_network and strength with strength following the existing descriptions in the proposal
Limiting the usage of such tags to: campground, trailhead, visitors center, and camp pitch
I know these will bring value to me, hopefully others find value as well.
I suggest removing âassumed.â Either the mapper has some evidence (e.g. they went to the site and tried to use their phone) that there is, isnât, or is poor coverage; or they donât have direct evidence. If there is no direct evidence, they should not use the cell_reception tag, just like if we donât know the speed limit for a road we donât use the maxspeed tag. If the data user/consumer wants to make some assumptions, they are free to do that.
Although atmospheric conditions impact reception, my experience has been if there is good coverage (cell_reception=yes) at some point, then most other times there will be at least limited coverage. If cell_reception=limited, the data user/consumer should know (and we should document), that there may be times when it is impossible to make a call, send a text, etc.
I know that other websites, such as Mountain Project (concerned with technical rock climbing) list whether a given site has cell coverage, e.g. Rock Climbing in Jurassic Park - Lily Lake, Estes Park Valley
They must think such information is valuable to their users (some of whom probably use trailheads, campgrounds).
hmmmm⊠I sense some frustration on your part. Hopefully we all can have a more constructive conversation going forward.
Thanks for the thoughts @tekim! I see your point, I agree assumed doesnât make much sense if this is an additional/optional tag to existing objects No reason to have basically an unknown tag when we could just not tag it, Iâll remove that!
Sorry to be awkward, but youâre going to have to go a lot deeper than that.
e.g.In Australia we have 3 major carriers (Telstra, Optus & Vodafone), who operate 4 networks (Telstra has separate Retail & Wholesale networks), together with ~30 MVNOs running off those various networks.
So for any spot, youâre going to have to, at least, say e.g. Telstra Retail=yes; Telstra Wholesale=limited, Optus=limited, Vodafone=no. & thatâs assuming that somebody using e.g. AldiMobile, knows that they are actually using the Telstra Wholesale network? If you assume they donât, then that should all mean 35 carriers being listed.
Once youâve sorted that, you also have to consider what phone anybody is using? Weâre standing side by side, but my older phone is working, while your brand new phone isnât, because Iâm still running on 3G via Telstra Wholesale, while your phone only has 4G / 5G off Telstra Retail, which is a much âbetterâ network, but with shorter range, and which doesnât have coverage in this particular location.
(Incidentally, over the next few years, the 3G network will be progressively shut down, so that side of it will all change!)
I understand what youâre saying, & donât disagree, but the actual machanics of adding this is pretty well impossible
I think this will be useful info, if it is made clear that it has outdoor leisure activities in mind, and is focussed on safety. I.e. I would very much support adding info to the map that might help me in an emergency situation to find a spot from where I am likely to be able to call 112/911 to ask for help. The tag could be added to backcountry campsites, but also to natural features such as mountain tops, springs, etc. and even to nodes that are specially created to tag them with this tag. We once camped at this site Way: 614876828 | OpenStreetMap in Kyrgyzstan and I remember it had a sign pointing to a nearly hilltop stating that âfor cell phone reception, go up thereâ. Iâd like to tag that hilltop with this tag.
In the case of one major carrier in the USA, 3G is fully dark (at least in major markets, I understandâŠthe more rural ones are on the chopping block by the end of 2023, Iâve been told, though this boundary is a bit squishy and has been pushed out more than once). But yeah, 3G is going the way of the dinosaurs eventually. Gee, and that particular (old) phone of mine still works great, and with its original battery, after something like 16 years! (Today: âNo Serviceâ as 3G is dark on that carrierâŠand yes because it was outrageously cheap <US$10/month, I did keep that number going for 15 years or so).
Again: WiFi âat a campgroundâ in OSM? Sure. Cell / mobile phone / commercial data reception? No, thanks. Too many changing, ephemeral, truly difficult-to-accurately map variables. And the carriers already make such maps (or they should, or their customers should ask for them). OSM doesnât need these data, itâs the ultimate nutty exercise of chasing oneâs tail. (And remaining unanswered: why would anybody do that? At this point, itâs a rhetorical question).
I might be miss-remembering here or someone already brought it up and I just didnât see it, but Iâm pretty sure you can tell cell tower range pretty easily based on the type of cell or whatever and the information is readily available to boot. You could probably then create a map of coverage based on that by adding to specific cell towers. Although itâs also contingent on which direction the cell is facing, but Iâm pretty you can find that out somewhere to. Say a cell has a 25 mile range on semi-flat terrain facing north east with a 30 degree radius or whatever. That could be turned into a fairly accurate map. Although with the caveat that itâs just an estimate, but whatever. Most of this is to begin with anyway.
@Adamant1 On the contrary, you can usually not produce a radio coverage map just from the antennas involved that would be useful for practical purposes. The most interest part is where coverage is bad or missing. Look at the attached picture to see how spiky it is with lots of isolated islands:
Propagation is influenced by the height of the base station antenna and the user, the frequency, transmit power, noise&interference, breathing, the Freshnel zone clearance and the bedrock material of the two. For built-up areas, you need to model most big trees and buildings to arrive at satisfactory results. It can thus be very expensive to produce, hence why proprietary models and data sets are used to simulate this by each provider. Some providers average out the simulation result for an easier overview or corporate interests, but weâve seen some who had shared the raw maps that was akin to the above example with the same blind spots, streaks and islands visible.
Oddly enough, we have advised against mapping exactly this kind of information in our Wiki years ago. Now I remember where Iâve seen it â Any tags you like - OpenStreetMap Wiki
What not to map
Entities in the OpenStreetMap database should relate to some geographical property or object with geographical qualities. For example, adding the location of a WLAN hotspot base station is considered acceptable, but tagging many nodes around it with the perceived signal levels is unwanted. Such information is better stored in a separate service.
Iâm not satisfied with the kind of response received on the mailing list. Their phrasing seems to imply that like half the people didnât bother to read the previous discussion or the proposal itself.
Why wigle or opencellid is not a good fit for this problem is that they only collect places with coverage from a certain provider, not places without coverage of the said provider. If they did even try to collect that, it would be tricky to implement it properly.
The same difficulty also applies to mapping this in OSM. Surveying in the affirmative is easy: if youâve got a signal, youâve got a signal. Surveying in the negative is not: is the lack of service because thereâs genuinely no service, because the towerâs temporarily overloaded and has contracted its service area to compensate, because your phoneâs frequency bands donât match up with the towerâs, or because youâre standing at the wrong end of the parking lot?
What is the overall purpose of this? Thereâs a vast section of my area that doesnât have cell reception and never will have cell reception, because the FAA wonât allow towers in the area due to the proximity of final approach for FLL.
Totally agree on that. And itâs not only 3G, 4G, 5G. Those are just the technologies⊠still you need a phone able to receive the specific frequency range. So even different phones capable of 5G using the same operator might have different quality of service.
Also we should keep in mind that the coverage is changing frequently, new towers get build kind of daily, old towers get upgraded from 3G to 5GâŠ
OSM-data will be always by far behind the actual published provider coverage data.
The only usecase I see is in very remote areas with no coverage but there is a dedicated hill/spot where you can go to call emergency response (which you can do in any kind of (gsm) network).
