How China's boundary data is compiled.

Most projects that need Chinese administrative boundaries end up assembling them from three or four partial sources, reconciling names by hand, and discovering six months later that two neighbouring provinces overlap by a kilometre. This site exists to make that a download instead of a week.

How it is built

The division list and the six-digit codes come from GB/T 2260, the national standard for administrative division codes. Geometry is compiled from public boundary data, then matched to that list: every polygon gets a code, and every code in the list gets a polygon. Where the source publishes something the list does not have — a forest reserve, a horse ranch, a directly administered county-level city — it is kept and labelled rather than deleted, because a hole in the map is worse than an unusual feature.

What gets checked

• Codes: format, uniqueness, and that each child's parent code exists at the level above.
• Names: English names reviewed by hand for the 368 second-level units and for the irregular county names; pinyin elsewhere, with the standard orthography for apostrophes and ü.
• Geometry: one feature per unit, closed rings, and a sane bounding box per feature.
• Formats: every file is re-opened after it is written — the shapefile with a shapefile reader, the DXF with a DXF reader, the XML against a parser — before it goes into the zip.

What it is not

It is not survey-grade. Boundaries are the published administrative lines at a generalised scale: right for analysis, mapping, planning and products; not right for settling a property dispute or grading a construction site.

Licence in one line

Use it for anything you like, including commercially, as long as you are not reselling the data itself as a dataset. The full text is on the licence page.