F-UJI score as indicator for the FAIRness of a dataset further increased
RADAR4Chem from FIZ Karlsruhe recently participated in the virtual support measure “FAIRness assessment challenge for datasets and semantic artefacts” of the EU project FAIR-IMPACT. As a result, the FAIRness of RADAR4Chem datasets was increased even further.
Evaluation tools and methods are increasingly being used to evaluate the FAIRness of research data. One of these tools is “F-UJI – Automated FAIR Data Assessment Tool”, a web service to programatically assess FAIRness of research data objects at the dataset level based on the FAIRsFAIR Data Object Assessment Metrics, which was developed as part of the FAIRsFAIR project.
With the help of mentors, FIZ Karlsruhe was able to further optimize the RADAR software during the support measure and thus increase the F-UJI score as an indicator of the FAIRness of RADAR4Chem datasets to values of up to 87%. The highest FAIR level “advanced” is achieved for the individual criteria “Findable” and “Accessible”. For the “Interoperable” and “Reusable” criteria, the FAIR level “moderate” was achieved for the majority of datasets.
Furthermore, the FAIR evaluation of various RADAR4Chem datasets once again demonstrated the importance of detailed metadata annotation using persistent identifiers for the FAIRness of a dataset. For example, if data providers do not link to a related resource (e. g. the DOI of a scientific article based on research data in the dataset), this leads to lower F-UJI scores. FIZ Karlsruhe is therefore considering integrating a “FAIRness check” with corresponding recommendations for data providers during metadata annotation in the future.
SPARQL Endpoint, Knowledge Graphs
On a daily basis, FIZ Karlsruhe builds a RADAR4Chem knowledge graph based on Schema.org vocabulary. The RDF triples of all RADAR4Chem datasets are available in turtle format (.ttl) from this URL: https://radar4chem.radar-service.eu/knowledgegraph
RADAR now also supports a SPARQL endpoint: https://www.radar-service.eu/sparql, with the help of which the knowledge graph can be automatically explored in a standardized way.