Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
ISO/IEC 25024:2015 — SQuaRE — Measurement of Data Quality
Quality measures for data quality characteristics including accuracy, completeness, consistency, and confidentiality
1. The Data Quality Measurement Landscape According to ISO/IEC 25024
ISO/IEC 25024:2015 addresses a critical gap in software quality management — the systematic measurement of data quality. While ISO/IEC 25023 focuses on system and software product quality, 25024 provides dedicated quality measures (QMs) for data, structured according to the 15 data quality characteristics defined in ISO/IEC 25012: accuracy, completeness, consistency, credibility, currentness, accessibility, compliance, confidentiality, efficiency, precision, traceability, understandability, availability, portability, and recoverability.
The standard introduces a two-perspective measurement approach: “Inherent” data quality, which concerns the data itself (its values, domain constraints, and relationships), and “System-dependent” data quality, which addresses how computer system components (hardware, system software) influence data quality. This dual perspective is essential for understanding the root causes of data quality issues — whether they originate in the data content or the enabling infrastructure.
When establishing a data quality program, prioritize the “Inherent” characteristics (accuracy, completeness, consistency) as they directly impact business decisions. System-dependent characteristics become critical when migrating or integrating data across different platforms.
| Data Quality Characteristic | Point of View | Example Quality Measure | Measurement Function |
|---|---|---|---|
| Accuracy | Inherent | Syntactic data accuracy (Acc-I-1) | X = A/B (syntactically accurate items / total items) |
| Completeness | Inherent | Record completeness (Com-I-1) | X = A/B (non-null data items / total items in record) |
| Consistency | Inherent | Referential integrity (Con-I-1) | X = 1-A/B (inconsistent references / total references) |
| Currentness | Inherent | Timeliness of update (Cur-I-2) | X = A/B (timely updated items / items needing update) |
| Confidentiality | Both | Encryption usage (Cnf-I-1) | X = A/B (encrypted values / values requiring encryption) |
2. Implementing Data Quality Measures Across the Data Lifecycle
A distinctive feature of ISO/IEC 25024 is its explicit linkage of quality measures to Data Lifecycle (DLC) stages and target entities. The standard identifies target entities such as architecture, contextual schema, data models (conceptual, logical, physical), data dictionary, data file, DBMS, RDBMS, forms, and presentation devices. For each target entity, specific properties are defined — attributes, elements, data items, data values, metadata, records, and information items — providing a precise framework for measurement.
This granular approach enables targeted data quality interventions. For example, during the Data Design stage, measures like “Conceptual data model completeness” (Com-I-6) and “Metadata accuracy” (Acc-I-6) help validate that the data architecture correctly represents business requirements. During Data Collection and Integration stages, measures such as “Referential integrity” (Con-I-1) and “Semantic consistency” (Con-I-6) ensure that incoming data maintains coherence with existing datasets.
Data quality measurement is not a one-time activity. ISO/IEC 25024 measures should be integrated into data governance workflows with regular monitoring cadences. A data quality scorecard built from standard measures provides management visibility into data asset health.
The standard provides 63 quality measures across the 15 characteristics, each documented with a unique identifier, name, description, measurement function, applicable DLC stages, and target entities. The measures are classified into three levels of use: “Highly Recommendable” (19 measures validated through practical use by large organizations), “Recommendable” (36 measures from innovative perspectives), and “For Reference” (8 measures from experimental research).
3. Engineering Design Insights for Data Quality Management
Implementing ISO/IEC 25024 in practice requires a systematic approach that integrates data quality measurement into existing data management frameworks. The measurement function pattern X = A/B (ratio of conforming items to total items) is used extensively, normalizing values to the range [0.0, 1.0] for consistent interpretation, where higher values indicate better quality.
Leading organizations deploy automated data quality monitoring using rules engines that continuously compute measures like Acc-I-1 (syntactic accuracy) and Com-I-2 (attribute completeness) against production data lakes. This transforms data quality from a reactive firefighting activity into a proactive, managed process.
A powerful technique from the standard is the use of outlier detection (Acc-I-4) for identifying anomalous data values that may indicate measurement errors, fraud, or systemic data quality issues. The standard describes both parametric (normal distribution) and non-parametric (quantile-based) methods for outlier identification, giving engineers flexibility based on their data distribution characteristics.
For organizations managing master data, the “Master data understandability due to metadata definition” (Und-I-3) and “Linked master data understandability” (Und-D-3) measures provide tools for assessing the effectiveness of metadata management. These measures directly correlate with the success of data governance initiatives and regulatory compliance efforts.
Q1: What is the difference between “Inherent” and “System-dependent” data quality?
A: Inherent quality concerns the data content itself (accuracy, completeness), independent of the system. System-dependent quality addresses how the computing infrastructure affects data usability (availability, portability, recoverability).
A: Inherent quality concerns the data content itself (accuracy, completeness), independent of the system. System-dependent quality addresses how the computing infrastructure affects data usability (availability, portability, recoverability).
Q2: How does ISO/IEC 25024 relate to ISO/IEC 25012?
A: ISO/IEC 25012 defines the data quality model (15 characteristics), while ISO/IEC 25024 provides the specific measures to quantify those characteristics. They are complementary standards within the SQuaRE measurement division.
A: ISO/IEC 25012 defines the data quality model (15 characteristics), while ISO/IEC 25024 provides the specific measures to quantify those characteristics. They are complementary standards within the SQuaRE measurement division.
Q3: Can ISO/IEC 25024 measures be applied to unstructured data?
A: The standard primarily addresses structured data in computer systems. For unstructured data, the measures for understandability, accessibility, and efficiency characteristics can be adapted with appropriate modifications.
A: The standard primarily addresses structured data in computer systems. For unstructured data, the measures for understandability, accessibility, and efficiency characteristics can be adapted with appropriate modifications.
Q4: What tools support automated data quality measurement per ISO/IEC 25024?
A: Commercial data quality platforms (Informatica, Talend, IBM) and open-source tools (Great Expectations, Apache Griffin) can implement many of the standard’s measures, particularly for accuracy, completeness, consistency, and timeliness.
A: Commercial data quality platforms (Informatica, Talend, IBM) and open-source tools (Great Expectations, Apache Griffin) can implement many of the standard’s measures, particularly for accuracy, completeness, consistency, and timeliness.
