place=city/town/municipality Revision

Finally, all 2,472 municipalities have been completely mapped as place=city/town/municipality, following the current guidelines after the discussion in

However, I’ve found that current guidelines are not perfect. I used to expect that the municipality that established the community would also be the municipality that is the urban area. This is sometimes not correct. Some urban municipalities have yet to establish the community (for example, Phrao (19.368630, 99.204272), Phanom Phrai (15.678812, 104.113322)), whereas some rural municipalities comprised of many distinct villages have already established the community (for example, That Thong (, Khun Thale ( Furthermore, finding official information on whether or not each municipality has already established the community is difficult. I admit that sometimes I’m unsure if it’s established or not.

So I’d like to revise the guidelines to make them more precise and easier to follow. According to OSM wiki, a town is “an important urban centre”, thus it may be a better idea to reflect the situation on the ground rather than strict to the administration status.

As a result, I suggest the following simple guidelines:

  • If more than 50% of the residential area* within the municipality boundary is connected as one settlement (no other area separating it), then tag it with place=city/town
    (A residential area includes anything from residential, commercial, public park, and other public urban facility, but does not include a gated private residential estate or an industrial estate. The presence of a natural obstacle, such as a river or a major highway, does not imply that the residential area is separated.)

  • If not, then tag it with place=municipality

How about this idea?

I finished editing the wiki and the map with this scheme.

Thanks for putting effort into this.
I see it problematic to get really objective definitions from the initially suggested rule without knowing for sure what the area is correctly tagged. Also whether it is separated by something or not might be a subjective question.

I hope that once tagged it will not require that many adjustments.I also hoe that we won’t end up in edit wars where people try to have their place rendered prominently in lower zoom levels by adjusting the place type.

I agree that there may be some issues, but this may be the best to distinguish the urban and rural municipalities in Thailand. At least until we have the better one.

In terms of edit wars, I’ve dealt with municipalities in OSM for a while and have rarely seen anything like that. The majority of these are misusing of the tag, such as changing some of them to place=suburb, which is quite strange and does not conform with any rules or sound reasoning. By the way, I believe one of the reasons it happens is that the previous criteria are so rigid and weird that we have to tag some of them as a town while no one calls them as a single settlement in real life.

Furthermore, I discovered some places where I tagged as place=municipality, some mappers then created a nearby new node with place=town, which I understand well because everyone calls it a town. Changing the guidelines to match more of the reality on the ground should be helpful.

Also, I believe it is acceptable for the “settlement” type of place=* to have some inconsistencies, as well as in other countries, and it is still in the amount that we can control. If there is any dispute over whether it is separated or not because it is unobvious, it is better than be forced to tag a group of many clearly distinct settlements as a single town just because the administrative “community” is established.

Some questions to above pictures

there are 2 place=neighbourhood tags, one with admin level 10 and one with admin level 11. Why is that?

for place=village should there not also be ‘if still exists’? If the communities are established and
the village has disappeared there should be no more village node.

BTW I agree with stephankn’s concerns. Local mappers will tag their own municipalities wrongfully with place=town. They will argue that the residential area is above the 50% limit. So will desktop mappers when they look at sat images. IMHO there is too much room for personal interpretation.

I have been adding/maintaining mostly remote/ethnic villages in northern Thailand, and
since I only know the local names (not the municipality boundaries), I tag them usually with:

  • place=village if the settlement includes facilities like a school, a temple, or a church
  • place=hamlet when there is only residential housing (typically less than 20/30 houses)

Does this still work with the current guidelines?

It is for the case where the village has not disbanded after the community is established, which is the most of cases, so that both can be mapped.

If the village no longer exists, either administratively or for addressing purposes, the municipality should be tagged as a city/town regardless, because we can’t find the names of each distinct village to map them. After reviewed, this case would be only 1 from the total of 2,472 municipalities, the Cha-am Town Municipality, which is currently mapped as a town. I’ve already added notes in the wiki about this case.

This was the case with the previous guidelines, where some towns were forced to be tagged as a place=municipality just because the community had yet to be established, and had no plans to be established. It looks weird, so some mappers changed it or added a new nearby place=town node. It may not follow the rules, yet it appears to make more sense to common users.

By the way, because all 2,472 municipalities are now completely mapped, there will be no new task for mappers to decide the tag, unless they want to change the ones that are already mapped. If time passes and the settlement expands, such that the residential area becomes larger and more like a town, the tag should be changed. There may be a debatable length of time, but at least it should be worth assuring that a clear rural municipality is not to be a place=town and a clear urban municipality is not to be a place=municipality.

Of course, the current guidelines are better suited to this case than the previous ones. For the previous one, if a village is inside the administrative boundaries of a municipality that is tagged as a town, it must be tagged as a neighbourhood, even if it is far apart. The revised guidelines solve this problem, thus if it is a distinct settlement, it can be a place=village/hamlet regardless of whether it is in a municipality boundary or not.

place=village is for an officially recognized “village” (muban), which is sometimes hard to tell, but your criteria appear to be valid in most circumstances.

A methodical approach to decide whether it is place=village or place=hamlet is to look for official sources.
You recently mapped villages in Mae Win Subdistrict. SAO Mae Win has 19 official villages.
These villages are tagged place=village. Other settlements in that area are place=hamlet.

Do you know if there is any online database available with all SAO <> villages relationships?
Doing a search per SAO going to be time-consuming.

You can check for villages in each municipality/SAO. It is located in the section “สังคม ศาสนา วัฒนธรรม” of each SAO’s page.

This is another site that could be easier to use. You can browse through province, district, and subdistrict. It is not 100% correct, but it should enough for validation purposes.

To avoid confusion with a community (admin_level=11, brown dot), I’d like to propose changing the village within the municipality area (admin_level=10, blue dot) to a place=quarter instead of a place=neighbourhood.

The only concern is that the wiki ( states that “For villages and small towns you might want to use this tag only if there is also smaller parts (neighbourhoods).”, however many of these municipalities don’t have smaller parts (community). If we are not taking this statement seriously (we can claim that in an ideal situation, the community should be established for all municipalities, though this is not the case in the real world), then using place=quarter is fine.

In theory using quarter instead of neighbourhood is also fine.
A problem could be that the neighbourhood (without admin_level) is with nearly 800 nodes relatively widely in use:

So will it bring additional benefit to use a different tag at all?
In practical life, a gated community with or without admin_level are hard to distinguish. So why make it harder for mappers than it needs to be? Tag both the same. If it has a formal approval, then add admin_level. Otherwise maybe a note stating that it doesn’t.

Our place for example is in “บ้า” (Ban) xxx. It has no official state as housing estate. For administrative purpose we had to visit the ผู้ใหญ่บ้าน (phu yai ban) in which responsibility area that housing estate is.

My bust guess had been that gated communities never have one. But in other thread you mentioned differently. So without other data sources or questioning the locals living there very hard to tell by just looking.

For the chart above that would then mean a note that such a housing estate would be also a neighbourhood.

Then having a housing estate as a neighbourhood would make sense. The admin_level doesn’t have to be added to the node the first time it is mapped.

P.S. I’m not sure if I explained it well. I propose the place=quarter to be used for a village (muban, Ban), not a community (chumchon). These two administrative entities are not the same. Both can coexist in a municipality. That is why they need to have different admin_level (10 and 11).

We have to be careful with the terms. A housing estate (unfortunately called village in Thailand) should never be called a village in OSM. When refering to a housing estate always call it as such.

What nitinatsangsit is proposing is to change a real Thai village (e.g. Mu 10 Ban Mamuang) within a place=town urban area from place=neighbourhood admin_level 10 to place=quarter admin_level 10.

The result would be

place=town Town name admin_level 7
place=quarter Ban Mamuang admin_level 10
place=neighbourhood Community Mamuang 1 admin_level 11
place=neighbourhood Community Mamuang 2 admin_level 11

Nitinatsangsit said above about place=quarter:
The only concern is that the wiki ( states that “For villages and small towns you might want to use this tag only if there is also smaller parts (neighbourhoods).”, however many of these municipalities don’t have smaller parts (community).

The wiki uses the word “might” and that is just a possibility.
So place=quarter looks fine even if there are no communities.

PS: This proposal changes nothing for a housing estate. They are tagged throughout Thailand as landuse=residential or place=neighbourhood without admin_level.

I completely agree, and this place=quarter will nicely prevent the confusion of having a place=neighbourhood inside another place=neighbourhood.

@nitinatsangsit I would suggest two improvements for the wiki table to prevent further confusion:

  1. add the place=quarter and its description together with the (see note) section
  2. add an extra column “No Tag” and include both place=hamlet for small settlements with no administrative status and place=neighborhood for housing estates not registered as communities.

Thanks nitinatsangsit and Mark_B_ for clarification.
I in deed was reading it wrong and thought it was a proposal regarding the gated communities.

Currently there are about 2000 place=village on admin_level=10 mapped. Is there any estimate how many of them would fall in the category to change to quarter?

I have never seen a housing estate registered as an official community of a municipality. I believe it does not exist.

for most place=village the admin_level is not mapped.

place=quarter is already used in Bangkok for a Khwaeng แขวง on admin_level 8. Is this a problem?

Good point. Then there would be no problem.

Agree. I’ll put it on the wiki.

Based on my experience, I guess more than 90% of them are still place=village.

For an example, you can see the Pak Nam Samut Prakan Town Municipality communities list, Most of them are housing estates.

I think it’s fine because the change has no effect on the Bangkok area.