Coverage Gap Audit¶
Coverage audit — instantiates Exhaustive Disjoint Partition Design
Reconciles what is actually assigned against the declared universe to surface the units that fall into no block — the exhaustiveness half of the partition promise, checked against real data.
Coverage Gap Audit tests the collectively exhaustive promise empirically. Take the declared universe, subtract everything currently assigned to some block, and whatever remains is a coverage gap — units the partition silently drops. Its defining move among its siblings is that it is a runtime, data-level check, not a design-time assertion: it does not build the partition or state its intended invariants, it finds where the live partition fails exhaustiveness, and it emits the gap record that stewardship acts on. Its mirror image is Overlap Detection Join, which checks the disjoint half; run together, the two verify both invariants against reality rather than against the drawing.
Example¶
A county's zoning map should assign every parcel of land to exactly one zoning district.[n1] The coverage gap audit joins the parcel registry — the declared universe, every taxable parcel on the roll — against the zoning layer and lists the parcels matched to no district: a freshly subdivided lot, a vacated right-of-way, an annexed strip that never got zoned. Each is a gap, and a consequential one — land with no zoning has no rules to apply, so a permit for it could slip through unreviewed.
The audit emits a record: parcel IDs, why each is unmatched, and when it first appeared. That record routes to the planning steward to assign or to the residual policy. The outcome is that "covers everything" is verified against the actual parcel roll — and the roughly forty unzoned parcels that had quietly accumulated over years are surfaced deliberately instead of discovered in the middle of a lawsuit.
How it works¶
- Anti-join against the universe. Conceptually,
universe LEFT JOIN assignments WHERE block IS NULL— the unmatched rows are the gaps. - Source the universe independently. The audit is only as trustworthy as the roster it treats as "everything," so that roster must come from outside the partition being checked.
- Classify gaps by cause. New units, retired blocks, timing lags, and genuine omissions get different dispositions.
- Emit a dated record and route it. Each gap is logged with its cause and handed to a steward or the residual policy, not silently closed.
Tuning parameters¶
- Universe source and authority — which roster counts as "everything"; the highest-leverage dial, because a too-narrow universe hides the very gaps you are hunting.
- Audit cadence — one-shot reconciliation versus a continuous monitor; continuous catches drift early but costs standing infrastructure.
- Gap-severity threshold — which gaps must block downstream use versus merely being logged for later.
- Reconciliation tolerance — how much timing or rounding mismatch is treated as noise before a discrepancy is called a true gap.
- Disposition routing — whether found gaps auto-assign, go to a steward, or drop into the residual bucket.
When it helps, and when it misleads¶
Its strength is catching the silent failure mode. A dropped unit produces no error and shows up in no downstream count — it is simply absent — so only a reconciliation against the declared whole reveals it. This is the check that keeps "exhaustive" from being an assumption.
It misleads exactly to the degree its universe is wrong. It can only find gaps relative to the roster it is given, so an under-declared universe yields a clean bill of health that is false. The classic misuse is to run it backwards — defining the universe to exclude the units suspected of being uncovered, so the audit dutifully reports "no gaps." The discipline that guards against it is to source the universe independently of the partition under audit, reconcile counts to a trusted external total, and re-run as the universe drifts — a passing audit against a stale or self-referential universe proves nothing.
How it implements the components¶
Coverage Gap Audit realizes the exhaustiveness-detection side of the archetype — the empirical test and its record, not the design of the blocks:
collective_exhaustiveness_invariant— it operationalizes this invariant and tests it against live data, turning "everything is covered" from a claim into a measured result.overlap_and_gap_audit_record— it produces the dated record of uncovered units and their causes, the artifact stewardship and the residual policy act on.
The disjointness/overlap half of that record is produced by Overlap Detection Join; it does not declare the universe or the intended blocks (MECE Partition Template), and it does not decide where a found gap goes (Residual Bucket Governance Policy, or Boundary-Case Triage Workflow when the gap is really a boundary case).
Related¶
- Instantiates: Exhaustive Disjoint Partition Design — it verifies the "covers everything once" promise against the data instead of assuming it.
- Consumes: MECE Partition Template supplies the declared universe scope the audit reconciles against.
- Sibling mechanisms: Overlap Detection Join · MECE Partition Template · Boundary-Case Triage Workflow · Block Membership Decision Table · Equivalence-Class Partition Derivation · Graph-Coloring Partition Assignment · Residual Bucket Governance Policy · Partition Refinement/Coarsening Review · Partition Crosswalk Table · Partition Change Notice · Stratified Partition Sampling Check
Editorial Notes¶
Form Classification¶
Form family: Assessment, Review & Assurance
Rationale: Coverage Gap Audit operates as a bounded evaluation of existing evidence or work that produces a finding or disposition because it reconciles what is actually assigned against the declared universe to surface the units that fall into no block — the exhaustiveness half of the partition promise, checked against real data.
Independent corroboration: The frozen evidence defines Coverage Gap Audit as 'Reconciles what is actually assigned against the declared universe to surface the units that fall into no block — the exhaustiveness half of the partition promise, checked against real data', so its operative form is Assessment, Review & Assurance.
Review outcome: Independent reviewer agreement; high confidence.
Origin Attribution¶
Primary origin: Earth Sciences
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Multi-domain
Rationale: Geographic information systems cohered topology audits that test polygon coverages for gaps and slivers; the generalized record-level audit combines that lineage with reconciliation, data anti-joins, and partition theory.
Related originating lineages:
- Accounting & Auditing — Reconciliation practice supplied comparison of an actual assignment ledger with an independently declared universe.
- Data Science & Analytics — Data-quality anti-joins supplied record-level detection of assigned and unassigned units.
- Mathematics — Partition theory supplied the exhaustiveness invariant that every universe member belongs to some block.
Review resolution: The mechanism source explicitly anchors coverage gaps and slivers in GIS topology, while its declared-universe anti-join is a deliberate cross-domain generalization.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Researched adjudication after independent review; high confidence.
Sources consulted:
Notes¶
"No gaps found" is only as strong as the universe definition behind it. A coverage audit that passes against a stale roster, or against a universe derived from the same partition it is checking, proves nothing at all. Pair it with a periodic reconciliation of the universe itself against an authoritative external count, so the denominator is trusted before any statement about coverage is.
[n1] In GIS, a coverage is a topologically clean layer whose polygons tile a space with no gaps or overlaps; the leftover unassigned fragments are called slivers. A coverage gap audit is the reconciliation that hunts for exactly those slivers. ↩