Subdistrict (tambon) level 8 boundaries import

Are there anybody that import level 8 subdistrict boundaries to thailand? I saw many site like esi have subdistrict boundaries map too is it possible for anyone to import those boundaries?

The import itself is not overly complicated. The big problem is that boundaries like these are not verifiable on the ground by the common mapper.

The data source we would use for the import has to be of proper quality and has to come with a license that allows us to use it within osm.

Do you know a source which would fit?

I may be wrong or not the first but I found it on Arcgis online I don’t know about the licence issue though

ArcGIS is proprietary and cannot be used without explicit written permission.

GeoBoundaries unfortunately only provides provinces and districts data, and Microsoft Open Data has already imported those with the appropriate permissions: Administrative boundaries of districts in Thailand - #11 by TaraV

Not aware of any other datasets we could use. Just curious, what purpose are you looking to reuse these boundaries for? Address lookup, trails discovery, or something else?

I’ve been trying to draw these borders myself by looking at images of subdistricts (Level 8). However, this method is not accurate at all. And as of now, I don’t think there’s a better method. The import method does sound fun, but where do we even get the data?

1 Like

I’ve reached out to the HDX / OCHA team to request permission to use their dataset. I’ll share an update once I receive a response:

If anyone comes across other dataset providers for subdistricts (ADM3), please add the link in this thread and indicate whether permission has already been requested.

This looks good, but how do you actually import it into OSM? I’m no expert, so of course I don’t know.

As someone who has worked on similar task, I can say that there is no accurate subdistrict map available for Thailand. Even official printed maps, such as the statute in the Royal Gazette, might contain errors of many kilometers. I doubt that even the government itself might not have it.

The Department of Lands may have the most detailed data because they know which land parcels belong to which subdistrict (albeit this is information from the time the title deed was granted, which may be a long time, like 100 years, when the subdistrict boundary/name is not the same as today). However, not all land in the country has a granted title deed. By the way, I’ve observed some local administrative organizations using the land parcel data to make maps that show their boundaries.

However, having a rough boundary is preferable to having none at all. When it comes to Thailand’s administrative boundaries, all maps, including those from Google, Apple, and others, are more imprecise than you think. If the data to import is unavailable to use with permission, the only quick method might be to roughly trace the map from the Royal Gazette. The outcome won’t be worse than other accessible commercial maps.

1 Like

You have a point! That’s exactly what I did for the subdistricts in Chon Buri. Of course, it’s still not accurate, but that’s probably the best option we have. I’ve already tried to make them as accurate as possible by making several edits.

1 Like

If I manage to get written approval to use the data, I can explain the process. Without permission, we’re not supposed to use it anyway.

Even the official paper boundaries for the country and provinces are sometimes not very accurate. But as you said, it’s still better than nothing, especially since it helps with address lookup and can be improved over time with local knowledge.

The tricky part with paper maps is that they use a different projection, so they need proper georeferencing and transformation before they fit well in a digital map. That’s doable, but still quite time-consuming.

1 Like

Yes, the Thai government usually uses the legacy Indian 1975 UTM (Zone 47 and 48, EPSG:24047 and EPSG:24048) on its maps. WGS84 UTM is being used by some of them (EPSG:32647 and EPSG:32648). This is difficult to solve by JOSM. They should be traced using QGIS or ArcGIS before importing them to OSM.

1 Like