Proposal Description

Official statistical agencies increasingly find themselves navigating the dual pressures of reducing bureaucracy while addressing the growing and evolving needs of users in a complex and rapidly changing world. Simultaneously, there is mounting pressure to lower costs for both reporting entities and statistical compilers, while optimizing the efficient use of existing resources. Innovative statistical products that aggregate granular data offer promising solutions to these challenges. These products are based on the "collect data only once" principle and follow the “multipurpose approach”, for instance using data for statistical as well as supervisory purposes. Statistical products based on granular data also have the potential to replace multiple existing aggregated surveys, thereby standardizing reporting obligations, minimizing reporting burdens, and improving efficiency.
For users, these new statistical products based on granular data provide greater flexibility, allowing for purpose-driven aggregation and easier integration with other datasets. However, they also present unique challenges. Their economic interpretation can be complex, particularly for non-experts, and reporting entities may encounter difficulties that necessitate targeted measures to ensure data quality. Furthermore, the technical utilization of these datasets is complicated by their large volume and intricate structure, while anonymization and compliance with data protection regulations become increasingly demanding.
This session will explore the experiences of official statistical authorities in managing granular data, focusing on data aggregation, compilation, and dissemination, as well as insights from users of these datasets. Presentations will highlight best practices, address challenges, and discuss the opportunities created by technical innovation in the field of official statistics.

Topics to Explore:
• Granular Data vs. Aggregated Data: Exploring the advantages and limitations of granular data as basis for official statistics compared to traditional aggregated data. This includes the role of granular data in modernizing statistical systems, meeting diverse user needs, and supporting multipurpose use while reducing reporting burdens (one example being AnaCredit). Case studies can highlight successes, challenges, and lessons learned.
• Flexible and On-Demand Statistical Products: Demonstrating how granular data enables the creation of flexible, on-demand statistical products that adapt to evolving real-world user needs. This includes addressing the technical and organizational requirements for producing customizable outputs.
• Best Practices for Aggregation: Discussing methodologies, algorithms, and frameworks for aggregating granular data into meaningful statistical outputs. Emphasis on ensuring consistency, reliability, and high data quality.
• Statistical Communication of Complex Data: Strategies for presenting aggregated statistics derived from granular data in an accessible and user-friendly manner. This includes addressing challenges in interpretation, ensuring transparency in methodology, and contextualizing outputs—especially when combining multiple datasets.
• Data Quality and Integration and the use of AI: Addressing issues such as data validation, error detection, and consistency across large, complex datasets. Exploring the use of machine learning and artificial intelligence for these purposes. Additionally, techniques for integrating multiple granular datasets, including overcoming challenges in data matching and harmonization.
• Balancing Data Utility and Privacy: Examining the balance between data utility and privacy, including innovative approaches to anonymization and compliance with data protection regulations. Highlighting how to develop innovative statistical products while adhering to legal and regulatory requirements.