+Trust / Provenance
DATA PROVENANCE: WHERE DID THIS FACT COME FROM?
Provenance is the record of origin and history behind data. It becomes especially important when tables are joined, several public datasets are reconciled, or an AI receives only the final output.
01 / Why it matters
A RESULT WITHOUT PROVENANCE IS HARD TO CHECK.
A value may look precise while its origin has been lost. Provenance keeps a chain between source and result: which source supplied the information, which version or date was used, and which transformations were applied.
In the DataAuthority framework, provenance is necessary but not sufficient. A perfectly traceable source may still have incomplete coverage, conflicting definitions or a limitation that makes the final claim unsafe.
- Origin: the source and dataset that supplied the fact.
- Time: when the source was valid or retrieved.
- Transformation: filters, joins, mappings, aggregation and normalization.
- Coverage: what the source actually observed.
- Limits: what the source cannot justify.
02 / Data Authority
LINEAGE TELLS THE PATH. DATA AUTHORITY ASKS WHETHER THE PATH SUPPORTS THE CLAIM.
That distinction is our own practical framing, not a universal industry standard. It helps us separate traceability from justification: knowing every step is valuable, but the final question remains whether those steps preserved the meaning needed for the intended use.