Opportunities and limitations of applied FAIR evaluation of data sets

Since their development, the FAIR principles have been met with broad acceptance in the scientific community. Tools based on various approaches are available to assess the FAIRness of individual data sets. These range from qualitative assessments based on questionnaires to automated quantitative measurements of fairness. As the FAIR principles are rather vaguely formulated, these approaches are based on individual, often differing, interpretations of the FAIR principles. In addition, the authors of the FAIR principles explicitly recognize the different implementations of FAIR within the various specialist communities. This makes it necessary to develop community-specific metrics and tests and to adapt FAIR assessment tools accordingly.

This diversity of methods for assessing FAIR is encouraging, as it sheds light on a variety of aspects of FAIR. However, this also sometimes leads to different, divergent results from these tools, which is difficult for users to work with. In addition, the measurement of FAIRness of individual datasets is heavily dependent on various technical implementations on the part of the data providers and their service providers. Numerous, possibly unintentional restrictions on the accessibility of datasets can influence or falsify FAIR measurements. 

In this presentation, we would like to report on our experiences with the applied FAIR assessment within this context. We will report on the further development of F-UJI, in particular our experiences with discipline-specific FAIR metrics and their implementation. Furthermore, we will discuss the limitations of FAIR measurements and try to delineate FAIR from aspects of data quality and accessibility and how to derive informative holistic assessments of datasets that include all these aspects in the future.