Thailand is not yet in the table of Tag:boundary=administrative. We should start to think about this. There is a proposal “Proposed features/capital” about tagging the
administrative places of regions.

There is a Wikipedia article Administrative divisions of Thailand.

Changwat, amphoe and tambon is quite clear for me. They are listed on many village signs from which I have collected this information and added to OSM. The Wikipedia article also mentions municipality (Thai: เทศบาล thesaban) and chumchon but is rather vague on this. Further information is needed.

Currently I’m “misusing” the tag “place” because I think the administrative level is more important than the population. In the countryside the places have less inhabitants and there would be a few cities and towns only and many villages.

  • place=city Provinzhauptstadt (Population 100.000 and over )

  • place=town Amphoe (Population between 10.000 and 100.000)

  • place=village Tambon (Population below 10.000)

  • place=hamlet Muban and below

We should try using the official classification.

For municipal areas it would be easy. It’s defined in the Thesaban Act (No 12)


Thailand divides its municipalities (thesaban) into three levels - cities (thesaban nakhon), towns (thesaban mueang) and townships (thesaban tambon).
The national capital Bangkok (pop. 5,713,288: Feb 2009) and the special governed city Pattaya (pop. 105,182: Feb 2009) are outside of these divisions.

Thesaban Nakhon (เทศบาลนคร)

currently 23 in Thailand:
* Nord-Thailand (ภาค เหนือ): Chiang Mai (เทศบาลนครเชียงใหม่), Lampang (เทศบาลนครลำปาง), Chiang Rai (เทศบาลนครเชียงราย), Phitsanulok (เทศบาลนครพิษณุโลก) und Nakhon Sawan (เทศบาล นครนครสวรรค์)
* Nordost-Thailand (ภาค ตะวันออกเฉียงเหนือ): Nakhon Ratchasima (เทศบาลนครนครราชสีมา), Khon Kaen (เทศบาลนครขอนแก่น), Udon Thani (เทศบาลนครอุดรธานี) und Ubon Ratchathani (เทศบาลนครอุบลราชธานี)
* Zentral- und Ost-Thailand (ภาคกลางและภาคตะวันออก): Nonthaburi (เทศบาลนครนนทบุรี), Pak Kret (เทศบาลนครปากเกร็ด), Ayutthaya (เทศบาลนครนครศรีอยุธยา), Nakhon Pathom (เทศบาลนครนครปฐม), Rayong (เทศบาลนครระยอง), Samut Prakan (เทศบาลนคร สมุทรปราการ) und Samut Sakhon (เทศบาลนคร สมุทรสาคร).
* Süd-Thailand (ภาค ใต้): Surat Thani (เทศบาลนครสุราษฎร์ธานี), Hat Yai (เทศบาลนครหาดใหญ่), Songkhla (เทศบาลนครสงขลา), Nakhon Si Thammarat (เทศบาลนครนครศรีธรรมราช), Trang (เทศบาลนครตรัง), Yala (เทศบาลนครยะลา) und Phuket (เทศบาล นครภูเก็ต)

Thesaban Mueang (เทศบาลเมือง)

it either needs to be a provincial capital, or have a population of at least 10,000
currently 118 towns

Thesaban Tambon (Thai: เทศบาลตำบล)

We don’t have the tag township (used in wikipedia). village would be closest in the sense that the population size is similar to what’s used in other countries for that tag.

For an area to qualify to establish as a Thesaban Tambon, it needs a gross income of at least 5 million baht, a population of 5,000, a population density of 1,500 per km², and the consensus of the population of the area.

1,456 Tambon as of August 15 2008

Muban (Thai: หมู่บ้าน)

As of 2008, there are 74944 administrative villages in Thailand[1]. As of the 1990 census, the average village consisted of 144 households or 746 persons.

as of today, the following 60 places are tagged as city:

“แม่ฮ่องสอน”;“Mae Hong Son”
“เชียงใหม่”;“Chiang Mai”
“สุราษฎร์ธานี”;“Surat Thani”
“กำแพงเพชร”;“Kamphaeng Phet”
“เชียงราย”;“Chiang Rai”
“นครศรีธรรมราช”;“Nakhon Si Thammarat”
“อุทัยธานี”;“Uthai Thani”
“นครปฐม”;“Nakhon Pathom”
“สุพรรณบุรี”;“Suphan Buri”
“นครสวรรค์”;“Nakhon Sawan”
“สมุทรสาคร”;“Samut Sakhon”
“สิงห์บุรี”;“Sing Buri”
“หาดใหญ่”;“Hat Yai”
“ปากเกร็ด”;“Pak Kret”
“ปทุมธานี”;“Pathum Thani”
“สมุทรปราการ”;“Samut Prakan”
“นครนายก”;“Nakhon Nayok”
“ปราจีนบุรี”;“Prachin Buri”
“นครราชสีมา”;“Nakhon Ratchasima”
“หนองบัวลำภู”;“Nong Bua Lam Phu”
“หนองคาย”;“Nong Khai”
“อุดรธาน”;“Udon Thani”
“ขอนแก่น”;“Khon Kaen”
“มหาสารคาม”;“Maha Sarakham”
“ร้อยเอ็ด”;“Roi Et”
“อำนาจเจริญ”;“Amnat Charoen”
“อุบลราชธานี”;“Ubon Ratchathani”
“นครพนม”;“Nakhon Phanom”
“สกลนคร”;“Sakon Nakhon”

Total places 1197:

Today I have created a first proposal. Please feel free to discuss it under this topic and to make corrections and enhancements directly at OpenStreetMap-Wiki: Thailand.

Edit: Enhanced and added version 2.

Thanks for adding this very interesting informations.

You are wrong, OSM is an encyclopedia. Not long time and Wikipedia will copy content like this from OSM, not only use our map.

I think the table in the current form is causing confusion.

We should keep two things separate.

First is tagging of areas. Changwat, amphoe, …
This classification is specified by the Thai administration. Our job would be to map these to the administrative levels used in OSM.
The proposal looks fine.

Another thing is tagging places. As a lot of names are used for areas as well as for places this can cause confusion.

They live side-by-side.

A thesaban mueang can have the same name as an amphoe.
The thesaban do administrative work for their region. It’s something like a “regional capital”.

Same as a town in western countries can have some districts (that have their own name), this is also true for Thailand.
The classification for the place tag would be city, town, village, hamlet. That matches well to the official Thai classification. See my post above.


Why is the table causing confusion? Is it the contents or the representation? What’s about having three rows for Thailand: English, Romanized Thai and Thai? Or how else could it be represented more clearly?

I can’t see the difference. Changwat, Amphoe, Tambon, Muban and Chumchon as well as Thesaban Nakhon, Thesaban Mueang, Thesaban Tambon are administrative areas defined and used by the administration.

I guess the problem is caused because some areas can be represented in OSM either by an area or by a node. Other areas can be represented by both: an area and a node. E.g. a city can be represented by its border lines (area) and a place (node). Where the node is set quite deliberately. A province could be represented also by an area and by a node but in OSM only by its border lines which mark the administrative area. There can be a node for its capital and a relation to bundle them. But the capital is not the province. The confusion is even greater because administrative areas are represented in some countries by a simple relation, in other countries by the multipolygon, the area of OSM.

I’ve added the OSM tags “place=city, town, village, hamlet” and moved Thailand to the top in the table. May be this might help to clarify. I don’t know a value for Key:place which can be used for Thesaban Tambon.

Edit: New version of table with separate rows for English, Romanized Thai and Thai. May be less confusing.

The tag “admin_level” is applied to ways and relations, usually in combination with “border=*”

The Thesaban I consider more appropriate for place= tags. and that is used on a node.

Thy not using place=village for Thesaban Tambon?
This is the definition for a place to get that status:
“For an area to qualify to establish as a Thesaban Tambon, it needs a gross income of at least 5 million baht, a population of 5,000, a population density of 1,500 per km², and the consensus of the population of the area.”

The size of the place (5000) fits well in the established “place=” tagging used in other parts of the world:

Exactly. That’s what I try to say.


I’ve fixed this for the places in the Northeastern region where I did this. 4 cities and 21 towns are now marked with “place=city” and “place=town” according to Wikipedia: List of Cities in Thailand.

The proposals how to mark capitals of a region seem to be stuck: Proposed features/capital and Talk:Relations/Proposed/Region.
To keep the information which places are the capitals of a Changwat, an Amphoe or a Tambon which I had hidden in the place tag I invented and established the tag “capital_level” which I will use until there’s an accepted solution. The same numbers as for Administrative levels are used:

  • “capital_level=4” marks the capital of a Changwat

  • “capital_level=6” marks the capital of an Amphoe

  • “capital_level=8” marks the capital of a Tambon

using capital_level should be fine for that. In case a different tag would be established to tag this it can easily be changed.

Is a capital always inclusive? The capital of changwat has amphoe mueang which also has a capital, usually the same…
…at least in Chiang Mai.


Wikipedia: Administrative divisions of Thailand: “Each of the provinces has one capital district (อำเภอเมือง, amphoe mueang), e.g. for Chiang Mai it’s Amphoe Mueang Chiang Mai. The exception is Ayutthaya Province, where both the province as well as the capital district have the full name Phra Nakhon Si Ayutthaya.”

This isn’t the case for Amphoe and Tambon. Some Amphoe have a Tambon and a capital with the same name. E.g. Ban Sida is in Tambon Sida which is in Amphoe Sida. Some Amphoe have a capital which is in a Tambon with another name. E.g. Kranuan is in Tambon Nong Ko which is in Amphoe Kranuan.

Can your tagging do this? What would the place-tag for Chiang Mai get in it’s capital_level? It’s capital of Changwat and Amphoe, right?

It might be easier if we would have already the boundaries, which we right now don’t have. A boundary relation of level x could include the capital node of that level.
Is the boundary-information by chance freely available? Did someone already ask or look for it?


I think it can do the job. The tagging for Chiang Mai would be and partially is already:

  • name=เชียงใหม่

  • name:en=Chiang Mai

  • name:th=เชียงใหม่

  • place=city

  • capital_level=4; 6

  • is_in:continent=Asia

  • is_in:country=Thailand (read as: is_in_admin_area_level:2=Thailand)

  • is_in:province=Chiang Mai (read as: is_in_admin_area_level:4=Chiang Mai)

  • is_in:district=Chiang Mai (read as: is_in_admin_area_level:6=Chiang Mai)

For all Changwat the capital is in the Amphoe with the same name, the Amphoe is in the Changwat with the name, the Changwat is in Thailand. Thus the "is_in" tags are unnecessary. In a strict database design duplicate information is to be avoided. Here they are for convenience of the mapper. If the "is_in" relation applies all over the world in the same way the editors can create this information "on the fly" from the spatial information and it's not stored again explicitely in the database.

Yes, capital can also be implemented with boundary relations. E.g.

  • boundary #1: name:en=Changwat Chiang Mai, capital=Chiang Mai, way1, way2, …

  • boundary #2: name:en=Amphoe Mueang Chiang Mai, capital=Chiang Mai, way1, way2, …

Both methods can do the job. They have advantages and disadvantages. When you look at the node "Chiang Mai" and it's tagged as above you see immediately that Chiang Mai is the capital of Changwat and Amphoe. If it's not tagged you have to search for and look through the boundary relations. If you look at the boundary first it's vice versa. I think a good solution would be to present the capitals to humans in both methods by editors but to store them only once in the database. And to allow capital without boundaries and boundaries without capital in the case you don't have both available.

Not long ago I learned about Global Administrative Areas. GADM is a Potential Datasource for OSM. User:Fsteggink asked already one year ago but didn’t yet get a reply from GADM.

GADM looks good at the first glance. You can download data for all or for individual countries. For Thailand it contains the boundaries down to Tambon including name and id. But I didn’t see mentioned the source and the date of the data for Thailand somewhere. There are big discrepancies for Tambon: Wikipedia: Tambon says 7255 Tambon plus 169 Khwaeng in Bangkok, GADM lists 5729 Tambon. E.g. for Mueang Khon Kaen: Wikipedia lists 18 Tambon, GADM lists only 12 Tambon.


I’m a native Thai citizen living in Khon Kaen. I’m still new to OSM. But as Willi2006 pointed me to this discussion, let me add what I can:

The capital Amphoe of a Changwat is not exactly with the same name. It’s with a leading “Mueang”. For example, the capital Amphoe of Khon Kaen is named “Mueang Khon Kaen”. And it’s usually called in short as “Mueang” in most occasions, except in some really official administrative documents. So, the name “Amphoe Khon Kaen” would sound plain wrong. “Amphoe Mueang Khon Kaen” is OK but still sounds like a lawyer’s or officer’s speaking, while “Amphoe Mueang” is more common.

That’s how the word “Mueang” is significant in the name. So, I think is_in:district value should be “Mueang Chiang Mai” in this case.

With this, I’d like to add more details to Thai administrative divisions.

There are 2 systems: the provincial administration and the local administration. The former is under the central government’s control, namely under the Department of Provincial Administration, Ministry of Interior. The latter is governed by locals themselves, under the support of Department of Local Administration.

The hierarchy is like this:

For provincial administration:
The Kingdom > Changwat (Province) > Amphoe (District) > Tambol (Subdistrict) > Muban (Village)

For local administration:

  • County (องค์การบริหารส่วนจังหวัด/Provincial Administrative Organization – อบจ./PAO), same as Changwat (Province) boundary

  • Municipality (เทศบาล/Tesaban) with 3 classes, under Amphoe (District) boundary:

  • City (เทศบาลนคร/Tesaban Nakhon)

  • Town (เทศบาลเมือง/Tesaban Mueang)

  • Township (เทศบาลตำบล/Tesaban Tambol)

[/*] [*]Tambol (องค์การบริหารส่วนตำบล/Tambo Administrative Organization -- อบต./TAO)[/*] [*]Special administration
  • Bangkok Metropolitan

  • City of Pattaya

[/*] [/list]

Chum Chon (ชุมชน/Community) is under Tesaban (City/Town/Township) governance. By definition, it’s not under provincial administration hierarchy, although its physical boundary can be inside a Tambol (Subdistrict) by coincidence.

For more information (mostly in Thai, sorry):

Hi Thep,

a warm welcome to the Thailand subforum and to the OSM community. I greatly appreciate native Thai citizen joining OSM. Thanks a lot for your contributions.

After reading your comments about “Mueang” I’ve checked the is_in:district values in the Northeast. I found 20 of about 300 had “Mueang” only and added the name: Changeset 4943542.

Your explanations of the Thai administrative divisions are very clear, comprehensive and helpful. Thanks a lot. I reworked the Wiki Draft accordingly and hope I got it right.