Certvas
Signal · Supplier-Issuer Demand v1

Public demand, mapped to the issuer.

Government tender award flow routed through the Certvas entity-resolution spine to a canonical issuer — so you can see which entities are winning public spending, how concentrated their buyers are, and whether they’re a listed issuer you can already screen on fundamentals. Every row carries its resolution confidence and point-in-time award window.

Grain: issuer (dataco_id) · point-in-time (award_date) · built on AF-TENDER (OCDS) × the entity spine · also via cv.signals.supplier_issuer_demand() and /v1/signals/certvas-supplier-issuer-demand-v1 · 4 provenance guarantees

Schema

Panel dictionary.

FieldTypeDescription
dataco_idstringCanonical entity id (the spine key the supplier resolved to).
issuer_name / jurisdictionstringResolved entity legal name and jurisdiction.
is_listed_issuerbooleanTrue if the entity also has AF-FUND fundamentals coverage. False = no fundamentals yet, not “unlisted”.
award_count / tender_countintegerAwards won, and distinct tenders behind them.
total_award_value / currencynumber / stringNominal award value (mixed-currency; sparse in OCDS — see caveats).
buyer_countintegerDistinct government buyers awarding to the entity.
top_buyer_share_pctnumberShare of awards (by count) from the single largest buyer — a concentration read.
first_award_date / last_award_datedateAward window bounds.
match_tier / match_scorestring / numberResolution confidence: exact (1.0) or fuzzy (≥ threshold).
observed_atdateWhen the award became knowable in Certvas gold (point-in-time anchor).

Honest coverage: the row set is bounded by spine resolution. South African tender suppliers are overwhelmingly private SMEs with no LEI, so only suppliers whose legal name resolves to a canonical entity appear — coverage is thin today and grows automatically as an SA company-registry source (e.g. CIPC) clears the licence register; the same resolver then lights up with no code change. total_award_value is nominal and mixes currencies (OCDS value coverage is sparse) — use award counts for cross-issuer comparison. Verify fuzzy matches before acting. See AF-TENDER and the provenance guarantees.