GMEOW Data Quality Module

• IRI: https://blackcatinformatics.ca/gmeow/slices/quality
• Tier: core
• Group: core

What This Slice Covers

This slice owns 11 terms and contributes 10 mapping or projection rows. Use it when its terms match the native fact you want to preserve; use the linkage tables to see how those facts leave GMEOW for consumer vocabularies.

Dependencies

• gmeow:slices/kernel
• gmeow:slices/observations

Consumers

• Data-quality claims (DQV/ISO 19157) over any slice's instance data; the coverage gates.

Local Map

Examples

Dataset Completeness

• Source: slices/core/quality/examples/dataset-completeness.ttl
• GMEOW terms: gmeow:Dataset, gmeow:Observation, gmeow:Organization, gmeow:QualityAssessment, gmeow:ScalarQuantity, gmeow:assessedEntity, gmeow:hasUnit, gmeow:methodComputationalModel, gmeow:name, gmeow:observationMethod
• External prefixes: xsd

# SPDX-FileCopyrightText: 2026 Blackcat Informatics® Inc. <paudley@blackcatinformatics.ca>
# SPDX-License-Identifier: CC-BY-4.0
#
# Worked example: data quality IS a measurement. A gmeow:QualityAssessment
# is a gmeow:Observation subclass, so a completeness assessment carries the SAME
# bundle as any other measurement: gmeow:assessedEntity (the feature, a sub-
# property of observedFeature), a gmeow:qualityDimension naming WHICH quality, a
# gmeow:observationMethod, a gmeow:vantage (who assessed), and an entity-valued
# gmeow:ScalarQuantity result (98.5% with its unit) rather than a bare number.
# Quality is not a special boolean flag — it is an observation like temperature.
@prefix gmeow: <https://blackcatinformatics.ca/gmeow/> .
@prefix ex:    <https://blackcatinformatics.ca/gmeow/examples/quality/> .
@prefix xsd:   <http://www.w3.org/2001/XMLSchema#> .

ex:dataset a gmeow:Dataset ; gmeow:title "Coastal Bird Survey 2026"@en .
ex:auditor a gmeow:Organization ; gmeow:name "Data Quality Audit Ltd."@en .

# --- The assessment: which entity, which dimension, by what method, from whom.
ex:assessment a gmeow:QualityAssessment ;
    gmeow:assessedEntity    ex:dataset ;
    gmeow:qualityDimension  gmeow:qualityDimensionCompleteness ;
    gmeow:observationMethod gmeow:methodComputationalModel ;
    gmeow:vantage           ex:auditor ;
    gmeow:observationResult ex:completeness .

# --- The result: a scalar percentage, unit-bundled like any measurement.
ex:completeness a gmeow:ScalarQuantity ;
    gmeow:quantityValue "98.5"^^xsd:decimal ;
    gmeow:hasUnit       <http://qudt.org/vocab/unit/PERCENT> .

Terms

Classes

Properties

Individuals

Linkages

• Rows: 10
• Projection profiles: -
• External vocabularies: dqv, oa, prov, wd

Guide

Quality — data-quality claims as ordinary observations

Slice: https://blackcatinformatics.ca/gmeow/slices/quality · tier: core The ISO 19157 dimensions, riding the universal Observation stack — quality is a claim like any other.

Data quality is usually bolted on as a separate report format. GMEOW refuses the bolt-on: a quality assessment is a reified observation about an entity, and so it reuses the universal Observation stack wholesale — the assessed entity is the observedFeature, the quality result is the observationResult (typically a ScalarQuantity), and the assessment protocol is the observationMethod. The result therefore carries unit, reference frame, determinacy, and provenance in the same bundle as every other GMEOW measurement (Principle 11), and the quality layer stays thin: one SubKind, two properties, one open dimension vocabulary.

The dimensions themselves are ISO 19157's — positional accuracy, temporal accuracy, thematic accuracy, completeness, logical consistency, topological consistency — plus lineage. W3C DQV and GeoDCAT-AP are aligned by reference (Principle 5); computing any quality metric is solver-layer work (Principle 12).

The core construct

gmeow:QualityAssessment

A reified assessment of the quality of an entity or dataset — a gufo:SubKind of gmeow:Observation, so everything the observations slice provides (vantage, method, result, frames) applies unchanged. The result is typically a scalar quantity (accuracy in metres, completeness as a percentage) or a categorical conformance statement. The EL-visible axiom guarantees every assessment assesses at least one entity; closed-world cardinality is SHACL's concern, never OWL's.

gmeow:assessedEntity

The entity whose data quality is assessed — the feature of interest of the quality observation, declared rdfs:subPropertyOf gmeow:observedFeature. That subsumption is the load-bearing move: a generic consumer can query "all observations about Alice" and get names, coordinates, and quality assessments without knowing this slice exists.

gmeow:qualityDimension

The dimension under which the assessment is made. Non-functional twice over: one report may evaluate several dimensions, and competing dimension classifications coexist rather than collapse (Principle 9).

The dimension vocabulary

gmeow:QualityDimension

An open value vocabulary of individuals — never subclasses (Principle 9). The seeds are the ISO 19157 set (qualityDimensionPositionalAccuracy, qualityDimensionTemporalAccuracy, qualityDimensionThematicAccuracy, qualityDimensionCompleteness, qualityDimensionLogicalConsistency, qualityDimensionTopologicalConsistency) plus lineage; a domain-specific "semantic consistency" or "usability" is added by minting a fresh individual, not a class.

gmeow:qualityDimensionLineage

The one seed that deserves its own note, because it is a bridge, not a duplicate: structural lineage already lives in the provenance module (wasGeneratedBy, wasDerivedFrom, ImportActivity). This dimension value marks an assessment that evaluates lineage completeness or correctness — provenance viewed through a quality lens, with the provenance graph itself untouched (Principle 4: one canonical source).

Reuse, not redefinition

The module mints no properties for axes that already exist — observationResult, observationMethod, and vantage come from observations; confidence and wasGeneratedBy from provenance; hasDeterminacy and hasUnit from the kernel; hasReferenceFrame from places; validFrom/validUntil from temporal. The reuse declaration in module.ttl is documentation of that fact, and the discipline is constitutional (Principle 4): a quality assessment's confidence is the same confidence every other claim carries, queryable the same way.

Boundaries

The slice records quality claims; it computes nothing. Conformance evaluation, accuracy statistics, completeness ratios, and the coverage gates that consume these assessments are solver-layer machinery (Principle 12). Alignment to W3C DQV (dqv:QualityMeasurement, dqv:Dimension) and GeoDCAT-AP is by reference and lossy projection — DQV has no standpoint, no determinacy, and no frame on its measurements, which is precisely what the Observation stack adds. Depends on kernel and observations; consumed by data-quality claims over any slice's instance data.