Adapting FAIR Evaluation to Photon and Neutron Facilities

Abstract

The FAIR principles have become essential in establishing transparent and trustworthy research practices. However the FAIR principles are guidelines indicating the features expected for data to be FAIR, and do not stipulate evaluation criteria. Consequently,there has been a proliferation of approaches to FAIR evaluation to substantiate claims for FAIR-ness, establish baselines, and measure improvement. Some approaches are focussed on FAIR-ness of individual datasets, others of repositories; some require extensive human evaluation, others use automation. However, within some scientific domains, data generation and management follow well-defined processes that result in datasets annotated with metadata and archived in repositories. Existing FAIR evaluation methods consider in less detail the contribution of the processes used in collecting and analysing data and how these enable FAIR-ness.

We describe the evaluation approach adopted for FAIR self-evaluation for Photon and Neutron Research Infrastructures (PaN RI’s). We review selected examples of existing FAIR evaluation frameworks designed to enable assessment at different levels, and outline four dimensions that characterise them. As no existing framework met our specific need to focus on FAIR workflows and processes inPaN RIs, it was necessary to select, combine, and adapt existing frameworks, and we developed an approach drawing heavily on the original FAIR principles, the RDA FAIR Data Maturity Model, and FAIRsFAIR’s CoreTrustSeal+FAIRenabling framework. Post-evaluation feedback from ExPaNDS partners indicated that they found the FAIR self-evaluation a useful and valuable exercise forunderstanding current levels of FAIR-ness at their facilities and for articulating what implementations they have in progress or planned to support FAIR in future.