H&S Enforcement and Injury Index

Every row is a verbatim sentence read from the named source's own page through our fetch service on the date shown, hash-pinned to the stored read (the sha of the page's visible text; the raw page is mirrored to R2).

2 of the 8 named sources that could be read published a usable figure; 6 published none and are recorded as not published, never filled in; 3 could not be read (blocked, dead or unreachable) and count nowhere.

The headline figures (median, minimum, maximum, quartiles) are over ONE figure per source for each breakdown it prices, in % (a source's median where its page yielded several accepted sentences for one breakdown), so a page that gives figures for several breakdowns gives one for each and N counts 6 figures, never sentences; 108 candidate sentence(s) were refused by a reviewer with the reason recorded.

The fuller derivation this dataset is chartered to grow into (not yet what the figures above are; they are each source's own published figure as read): one series per industry per year: enforcement notices served, prosecutions concluded, the regulation most cited, and the injury and ill-health rates from the national statistics release, with the fine distribution where the register publishes it; the derivation nobody publishes is the JOIN - enforcement volume per REGULATION mapped onto the statutory-assessment table, so a hazard row carries how often failing it is actually enforced, and the per-industry rate that turns 'you must assess' into 'here is what happens when you do not'; changed-since on every statistical release

108 candidate sentences on these pages were refused by a reviewer; each refusal and its reason is recorded in the review ledger.