China city place datasets: 386 datasets, 786 files, 13,482 records
One page per city or region, each one a small dataset with a count, a definition and a date, with the Chinese name kept beside the English one so the translation can be checked rather than trusted.
The numbers
| Value | What it counts | Source / definition |
|---|---|---|
| 386 | city and region datasets | Directories under /china/ that contain a page, counted from disk on 2026-09-26. The 387th directory holds the raw data files and is not a dataset. |
| 786 | JSON files served under /china/data/ | Counted on 2026-09-26: 390 hold place records, 384 hold the district-level breakdown for the same cities and 12 are companions for food, scenic spots and opportunity notes. |
| 13,482 | place records in the record files | Summed from the 390 record files on 2026-09-26, one record per place. Includes lists that are not attached to a single city, such as the country-wide and world lists. |
| 13,063 | records printed by the dataset pages themselves | 381 of the 386 pages state a count in their own description and those counts add up to 13,063. The gap against the file total is explained below rather than averaged away. |
What one record is
One record is one place in our own place database: a name in Chinese and English, a category, a rating, a coordinate and the district it sits in.
Records
- What is counted. One record is one place: name in Chinese, name in English, category, rating, coordinate and district. Rows without a usable coordinate are dropped rather than placed at a city centre, because a wrong dot is worse than a missing one.
- Why two totals appear. The record files add up to 13,482 places while the pages print 13,063, because five of the 386 pages do not print a count and six record files — including a country-wide list and a world list — do not belong to a single city page. We publish both figures instead of the friendlier one.
- Where the raw files are. Each page is built from JSON served under /china/data/, and the same directory is public. The files are the dataset; the page is a readable view of one file.
- Translation policy. The Chinese name is the record and the English name is a convenience. Where the English is wrong, the Chinese name still identifies the place, and the fix is a correction to a row rather than a rewrite of a page.
- Licence. CC BY-NC 4.0. Attribute InfoGap and link to the page you took a specific record from, which is also the honest way to cite a dataset that is rebuilt.
Coverage is not uniform and is not meant to be: a well-mapped city produces a long list and a small region produces a short one. A short list is evidence about mapping effort, not about the place.
Data & methods
- Source. Our own place database, cleaned and translated by us. The records are served as JSON next to each dataset page; no third-party page is scraped and re-served as ours.
- Definition. One record is one place in our own place database: a name in Chinese and English, a category, a rating, a coordinate and the district it sits in.
- Last updated. 2026-09-26
- Licence. CC BY-NC 4.0 — attribution: InfoGap (infogap.store), link to this page.
Where to go next
- Data pages: https://infogap.store/china/
- Method pages: https://infogap.store/method/how-we-build-a-city-dataset/
- Hub: https://infogap.store/data/
- Machine-readable: registry.json