Proposal Description

Statistical Validation and Quality Assessment of Alternative Data Sources for Official Statistics: Emerging Methodologies for Trustworthy Data Ecosystems
The increasing availability of alternative data sources is transforming the production of official statistics. Citizen-generated data, administrative records, Statistical Business Registers (SBRs), geospatial information, machine learning applications, and AI-assisted data products offer unprecedented opportunities to improve the timeliness, granularity, and relevance of statistical outputs. However, their integration into official statistical systems raises important methodological questions regarding data quality, validation, comparability, bias, transparency, and fitness-for-use.
This Invited Paper Session will explore emerging statistical methodologies and practical approaches for validating and assessing the quality of alternative data sources used in official statistics. The session brings together experts from National Statistical Offices, the United Nations system, geospatial science, and the Young Statisticians Network to examine innovative frameworks and applications across diverse data ecosystems.
Presentations will cover the development of a Statistical Quality Index for assessing the fitness-for-use of alternative data sources in official statistics; validation frameworks for Citizen-Generated Data; methodological approaches for integrating alternative data into Statistical Business Registers; machine learning applications for statistical validation; and quality assessment of geospatial data for official statistics and Sustainable Development Goal (SDG) monitoring. Collectively, these contributions address one of the most pressing challenges facing modern statistical systems: how to leverage emerging data sources while maintaining scientific rigor and quality standards.
Beyond methodological innovation, the session will also examine the broader implications of alternative data integration for trust in official statistics. As statistical systems increasingly adopt AI-assisted production methods and non-traditional data sources, maintaining public confidence in official statistics becomes more critical than ever. Discussants will reflect on issues of transparency, accountability, governance, and public trust, including the role of official statistics in supporting informed decision-making and democratic societies.
Particular attention will be given to the proposition that “Democracy Dies in Darkness Without Official Statistics,” highlighting the essential role of trustworthy, independent, and high-quality official statistics in promoting transparency, evidence-based policymaking, and democratic accountability.
The proposed session aligns closely with the World Statistics Congress theme on statistics and data science in the era of AI. By linking methodological advances in data validation and quality assessment with broader questions of trust, governance, and public value, the session aims to contribute to the development of trustworthy data ecosystems and resilient official statistical systems worldwide.
The session will further explore the development of standardized quality assessment frameworks that enable National Statistical Offices to evaluate the fitness-for-use of alternative data sources across different statistical domains. Particular attention will be given to methodological issues relating to uncertainty quantification, bias detection, comparability, reproducibility, and integration within existing quality assurance frameworks. Through presentations and discussant reflections, the session will identify practical pathways for balancing innovation with the professional principles that underpin official statistics, including relevance, impartiality, transparency, and scientific independence.