Assessing FAIR data principles against the 5-Star open data principles

Abstract

Access to biomedical data is increasingly important to enable data driven science in the research community. The Linked Open Data (LOD) principles (by Tim Berner-Lee) have been suggested to judge the quality of data by its accessibility (open data access), by its format and structures, and by its interoperability with other data sources. The objective is to use interoperable data sources across the Web with ease.

The FAIR (findable, accessible, interoperable, reusable) data principles have been introduced for similar reasons with a stronger emphasis on achieving reusability. In this manuscript we assess the FAIR principles against the LOD principles to determine, to which degree, the FAIR principles reuse LOD principles, and to which degree they extend the LOD principles. This assessment helps to clarify the relationship between both schemes and gives a better understanding, what extension FAIR represents in comparison to LOD.

We conclude, that LOD gives a clear mandate to the openness of data, whereas FAIR asks for a stated license for access and thus includes the concept of reusability under consideration of the license agreement. Furthermore, FAIR makes strong reference to the contextual information required to improve reuse of the data, e.g., provenance information. According to the LOD principles, such meta-data would be considered interoperable data as well, however, the requirement of extending of data with meta-data does indicate that FAIR is an extension of the LOD (in contrast to the inverse).