How I verify what I publish
Methodology
Who reviews this site
HanziCoverage is reviewed by the Hanzi Coverage editorial team. We also build Wordbrush, an iOS spaced-repetition Chinese-vocabulary app, and we maintain the word frequency and HSK coverage datasets both projects run on. That is the real, first-hand experience this page stands on: building Chinese-learning tools and keeping the underlying word data current, not a degree, a license, a native-speaker background, or a teaching credential. We do not claim any of those, and we do not need to: the authority here is the sourcing and computation process below, which anyone can check.
How data is sourced and computed
Every coverage figure on this site is computed from an openly-licensed Chinese
corpus (subtitles, news, encyclopedic, literary, or classical text), matched word by
word against the official HSK vocabulary lists. The corpus, its license, its source
URL, and the match rate are published next to every figure, in the per-page footnotes
and in the PROVENANCE.md files shipped alongside the data in this site's
own repository. I never hand-type a coverage percentage; every number is generated
by a build script from the underlying token counts, so a correction to the source data
moves every page that cites it on the next build.
The statistics page is this site's own original data study: an aggregate comparison across every reading register, computed at build time from the same sourced dataset, never hand-typed.
This site's freshness policy re-checks published facts every 90 days. A page that falls out of that window is queued for re-verification rather than left to go stale.
Last full verification
The newest verified fact on this site is dated . This date is computed from the site's own content and coverage dataset, not hardcoded, and moves automatically as guides and data get re-verified.
My connection to Wordbrush
I built Wordbrush, and I use it every day to learn Chinese myself. So when I recommend it on the recommended-app page, I am recommending my own work, and I would rather say that out loud than bury it. The analyzer, the difficulty grader, and the coverage matrix are free and always will be, whether or not you ever install anything.
Corrections
Spot something wrong? Tell me what page, what figure, and what it should say instead. You can reach me through the Urban Algorithm contact page. HanziCoverage is a project of Urban Algorithm LLC.