Report on data homogenisation for high-value datasets

Following the European Union’s open data policies, Directive 2019/1024 (EU), also known as the open data directive, defines the concept of high-value datasets (HVDs) as datasets whose reuse is associated with significant benefits for society and the economy that public sector bodies will have to make available under very reuse friendly conditions. On 21 December 2022, the EU adopted an Implementing regulation lying down a list of specific high-value datasets in each of the six main categories identified in the directive: geospatial, Earth observation and environment, meteorological, statistics, company and company ownership, and mobility. The Regulation must be implemented by June 2024. Starting in February 2025, Member States will have to report to the Commission every two years on specific high-value datasets available at the Member State level, including links to the licensing conditions and the APIs. Where available, Member States should also report on their guidance documents, data protection impact assessments, and possible temporary exemptions from the no-charging rule. Effectively used interoperable metadata will play a crucial role in the findability of the datasets by re-users and for easy reporting by Member States. The identification of these specific HVDs is a relevant step towards improving cross-border interoperability and dataset reuse. However, the high level of heterogeneity in the data formats, structures and levels of granularity in which these datasets are available, the lack of good-quality metadata about datasets (including metadata that allows one to determine whether a dataset can be considered a HVD or not) and the lack of interoperability of published datasets means that companies, entrepreneurs, researchers, data journalists, policymakers and citizens will still find it difficult to make use of these important resources, especially when using data from several countries. In this study, we propose a methodological approach that facilitates the identification and homogenisation of HVDs. We implement this approach using selected examples of HVDs from three distinct categories, showcasing its adaptability for different types of HVDs.