Overview
ISO/IEC TR 5259-6:2026 sets a global framework for the visualization of data quality within analytics and machine learning (ML) applications. Published by ISO and developed by ISO/IEC JTC 1/SC 42, this technical report provides practical guidance on leveraging visualization methods to effectively assess and communicate data quality measures. The framework is designed to aid a diverse range of stakeholders - including AI producers, providers, developers, users, and regulators - in understanding, evaluating, and improving data quality across an AI or ML data management life cycle.
The standard enhances trust and transparency in artificial intelligence by making data quality results tangible, actionable, and accessible. By integrating visualization into data quality management processes, organizations can drive clearer insights, identify data issues, and make informed decisions that elevate the performance and reliability of AI and ML systems.
Key Topics
-
Data Quality Management Life Cycle
Aligns visualization practices with every stage of the data quality management life cycle (DQMLC), supporting continuous validation and verification.
-
Stakeholder Perspectives
Supports the different viewpoints of AI system stakeholders, including those involved in the creation, deployment, use, and regulation of AI applications.
-
Dataset Properties and Context
Emphasizes understanding dataset characteristics-statistical properties, source, structure-which influence the choice and impact of quality visualizations.
-
Data Quality Models and Characteristics
Outlines common data quality characteristics such as accuracy, completeness, consistency, relevance, and timeliness, based on ISO/IEC 25024 and other standards.
-
Data Quality Measures and Assessment
Encourages quantifying and visualizing measures such as semantic accuracy, attribute completeness, and risk of inaccuracy to support informed evaluation.
-
Visualization Methods and Considerations
Discusses practical visualization techniques (e.g., bar charts, box plots, radar charts, dashboards) and best practices for presenting data quality.
Applications
-
AI and ML Development
Visualization frameworks are critical for AI/ML teams to communicate data quality findings throughout the development cycle, from data preparation to model validation.
-
Regulatory Compliance
Helps organizations demonstrate data quality controls, supporting compliance with industry regulations and guidelines.
-
Stakeholder Communication
Provides stakeholders-including technical, business, and regulatory audiences-with clear, graphical insights into data quality, supporting trust and transparency.
-
Data Quality Reporting
Streamlines the documentation of data quality management processes by integrating visual summaries, making reporting more engaging and effective.
-
Exploratory Data Analysis
Enables identification of common data quality issues, such as missing values, outliers, or inconsistencies, facilitating faster and deeper exploratory analysis.
-
Risk Assessment and Decision Support
Allows users to visualize and act upon data quality risks, supporting more robust, evidence-based decision-making in AI deployment.
Related Standards
The visualization framework in ISO/IEC TR 5259-6:2026 is closely connected to other core standards within the ISO/IEC 5259 series and supporting standards in AI and data quality:
- ISO/IEC 5259-1:2024: Defines general concepts, terminology, and examples for data quality in analytics and ML.
- ISO/IEC 5259-2:2024: Details data quality models and characteristics.
- ISO/IEC 5259-3:2024: Outlines the data quality management life cycle, synchronizing with visualization processes.
- ISO/IEC 22989:2022: Provides foundational concepts and terminology for artificial intelligence.
- ISO/IEC 25024 and ISO 8000 series: Specify detailed data quality characteristics and assessment methods.
- ISO/IEC 5339:2024: Describes stakeholder roles and perspectives in artificial intelligence applications.
- ISO/IEC 23751: Covers dataset properties and data sharing in cloud computing and distributed platforms.
Adhering to these standards ensures that organizations can implement best practices in data quality visualization, supporting innovation and enhancing the trustworthiness of AI and ML solutions.