Dataset
A metadata inventory of the pages on this site: 275 pages (176 Simplified source pages, 99 language-branch pages, grouped into 178 clusters), 600,900 characters of body text, 467 images. The snapshot is dated 24 October 2026 and matches the pages published on the site at that moment. Issued as a single CSV, for searching, counting and machine reading.
1. What is in it
One page per row, thirteen columns. The columns carry metadata, not the body text — the text stays on the page; this file only tells you which pages exist, what each one is, how long it is, and whether it has images and sources.
| Field | Meaning |
|---|---|
czid |
Entry identifier, in the form cz:<path with trailing slash>; the root page is cz: |
title |
Page title (a single-language string, one per row) |
slug |
Last segment of the URL path; a change requires a 301 |
url |
Canonical address of the page |
section |
Section the page belongs to. On translated pages this gives the Chinese section name of the source page; the page’s own machine-readable block gives the section name in that language |
pageType |
What the page itself is: Article / CollectionPage / WebPage / Dataset / FAQPage / DefinedTermSet |
subjectType |
What the page is mainly about: Person / Place / TouristAttraction / Event / czns:* / Thing |
lang |
Language branch: zh-hans / zh-hant / en / ja / ko |
cluster |
Cross-language correspondence: all language branches of one source page share a value (the source slug) |
chars |
Body length in characters (tags and whitespace removed) |
images |
Number of <img> elements in the body |
has_sources |
Whether there is a “Sources” section: yes / no |
modified |
Save time as recorded by WordPress (ISO 8601). The “last revised” line at the top of a page is separate and changes only on substantive edits; the two may differ |
2. Download
- chaoshanwiki-pages.csv (276 lines: 1 header + 275 data rows; UTF-8, comma-separated)
There is no separately hosted JSON version: the platform whitelist for this site does not allow .json or .txt uploads and rejects them. The same data can be taken from the site’s own REST interface — https://chaoshanwiki.com/wp-json/wp/v2/pages?status=publish&context=edit&per_page=100&page=1 — page by page until a request returns nothing. That interface returns raw WordPress fields, which you have to map onto the columns above yourself.
The CSV goes into spreadsheets and databases easily; the REST interface suits programs that read directly, at the cost of mapping the fields. The file is unencrypted, uncompressed and unrestricted.
3. Two layers of type, and the cross-language key
Two columns are easily misread, so they get a paragraph each. First, pageType and subjectType are not the same thing: the first says what the page itself is (Article, CollectionPage, WebPage and so on), the second says what the page describes (Person, Place, czns:IntangibleHeritage and so on). A page on Chaozhou opera is an Article whose subject is czns:IntangibleHeritage — both hold at once and should not be flattened into one column. The values are not invented here: they are copied from each page’s own JSON-LD block, so when a page changes they change with it on the next issue.
Second, cluster records the cross-language correspondence. The Chinese, traditional-Chinese, English, Japanese and Korean pages of one piece of content share a cluster value (the source slug), so a machine can treat the versions as versions of one object rather than as unrelated entries. The 275 pages currently fall into 178 clusters.
Full definitions of the types and fields are in the namespace.
4. Coverage
The dataset covers published pages, including the Simplified source pages and the zh-hant, en, ja and ko branches. Current counts:
| Item | Count |
|---|---|
| Pages | 275 |
| …of which Simplified source pages | 176 |
| …of which language-branch pages | 99 |
| Body characters (tags and whitespace removed) | 600,900 |
| Body images | 467 |
| Pages with a “Sources” section | 199 |
275 is the actual number of published pages at the snapshot of 24 October 2026, the home page included. Pages published after that date enter the file only on the next issue, so the two will drift apart in the meantime.
5. What is left out
Four things, stated plainly. Full body text is not included — that would amount to breaking the site up into a dataset, which conflicts with a reading format that carries captions, breadcrumbs and revision dates. Image files are not included either; only the count, with each image’s licence status marked in its caption. Unverified statements are not included, so the file has no “claim” field; claims are in the fact sheet. And only published pages are included: drafts and private pages stay out. This issue is every page published on 24 October 2026; later pages are added at the next issue.
6. Licence and citation
The dataset is released under CC BY 4.0: copy it, redistribute it, adapt it, with attribution to Chaoshan Wiki and a link back to the data page. Citation:
Chaoshan Wiki: “Chaoshan Wiki page and entry inventory”, version 2026-10-24, https://chaoshanwiki.com/dataset/
The body text is licensed CC BY-SA 4.0, which is not the same as the dataset: the dataset is CC BY 4.0 because it is a factual inventory with no expression in it. Copyright in the images is reserved and commercial use requires a separate agreement — see Copyright and licensing.
7. Updates and the three dates
The file is updated as the site is published, with no fixed schedule promised. Each file carries its version date at the top, and the field table changes when the field structure changes. Previous issues are not kept — write in and explain what you need if you want a historical version.
Three dates are not the same date, and are recorded separately: the date the whole table was checked is the version field (currently 2026-10-24, the snapshot date); the per-page save time is in the modified column; the check date of a single fact is the verified field in the fact sheet. Take whichever you need, and do not pass the version date off as the check date of a particular fact.
8. Neighbouring files
The namespace defines the cz and czns prefixes, the two layers of type and the fields; the fact sheet gives one citable line per assertion; Sources and evidence sets out the rules for citation.
Home › Machine-readable › Dataset
Last revised: 3 November 2026