Collect
Public, no-login sources are collected into the backend evidence layer.
Transparency
Every published figure should be traceable to a public source, a documented transformation, an OFO occupation match and a dated evidence snapshot. PathLens does not treat a modelled or proxy number as an observed fact.
Public, no-login sources are collected into the backend evidence layer.
Titles, locations, salary periods and source fields are standardised without overwriting the original observation.
Records are mapped to the six-digit OFO occupation code with a recorded confidence level.
Duplicates, ambiguous matches, weak samples and exceptions are identified before publication.
Once per month a dated evidence snapshot is created and held for admin review.
OFO is the canonical occupation spine. Six-digit occupation metrics are only published where the record can be defensibly mapped to a six-digit OFO code.
Official demand signals such as OIHD are kept separate from observed public vacancy counts. A vacancy count describes the public advertisements observed by PathLens; it is not the total number of vacancies in South Africa.
Worker earnings, advertised-market salary aggregates, public-service salary scales and individual advertised salaries are stored as different evidence classes. They are not silently blended into one salary number.
The raw salary and pay period are retained. Where an amount is converted to a monthly or annual comparison value, the original value remains stored alongside the derived value and the conversion method.
Repeated vacancy notices are identified using identifiers such as department or employer, reference number, normalised title, location, dates and description similarity. Public metrics count unique opportunities where possible rather than raw pages discovered.
A = exact or validated alias. B = strong title and task match. C = probable and requires review. D = ambiguous and requires review. X = unmapped. Only A/B records should feed automatic six-digit occupation metrics.
A figure is not strengthened by hiding a small sample. Salary and vacancy outputs should carry the observation count, source period and geography. If the evidence is insufficient, PathLens says so.
The evidence pipeline is scheduled for the first day of every month. Each run produces a new snapshot with successes, exceptions and coverage statistics. New official publications are incorporated when available.
A completed monthly run is marked Ready for review. It does not automatically replace the public dataset. An administrator approves the snapshot before it becomes the public evidence layer.
Source freshness
Official statistical sources used by the Statistics hub.
Last completed check: 2026-10-01T11:06:40.526673+00:00 Latest completed source check found no changed source fingerprints.Institutions, NQF, professional bodies, tax/ID/inflation reference sources.
Last completed check: 2026-10-01T11:07:00.001156+00:00 One or more official sources could not be fetched. 4 changed · 3 failedFunding-programme rules, policy and eligibility source monitoring.
Last completed check: 2026-10-01T10:32:06.642830+00:00 One or more official sources could not be fetched. 0 changed · 8 failedOfficial university fee-source fingerprints and review-gated extraction candidates.
Last completed check: 2026-10-01T11:21:32.648210+00:00 One or more official sources could not be fetched. 0 changed · 6 failedSource register
PathLens is a reference layer, not a census of every vacancy or salary in South Africa. Its figures describe the evidence that could be responsibly observed, mapped and documented from the stated sources.