Speakers
Description
The domain agnostic metrics adopted by FAIR data assessment tools tend to penalise metadata aggregators, like the CESSDA Data Catalogue (CDC). This became apparent during the work done for the ‘Bulk FAIR assessment of the CESSDA Data Catalogue using the F-UJI API’ (as presented at EDDI2022).
Building on that work, FAIR scores were generated by the F-UJI and FAIR EVA tools for some of the records held in the CDC and the same records at source. Then they were analysed to determine the relative differences and underlying causes on a criteria by criteria basis.
In order to do this, the script previously used to run the bulk assessments was modified to allow it to run different tools against different sources of metadata. The results were stored as ElasticSearch indices, and a Kibana dashboard provided a visual representation of the relative differences in the scores.
The results show that the lower scores for aggregated metadata records are due to the definition of the FAIR criteria rather than their implementation by the tools. Which begs the question, ‘is it OK that aggregated metadata is, by definition, less FAIR, or should the definition be changed to level the playing field?’