FAIR Data at Scale to Enable Open Science

Since they were published in 2016, the FAIR Principles for data (Findable, Accessible, Interoperable, and Reusable) have garnered much attention. Yet there is little consensus on how to put them into practice. There are at least nine formal assessment methods advocated by different organizations and dozens if not hundreds of reports and papers describing how different disciplines and organizations have endeavored to make their data FAIR. The FAIR principles have focussed conversation, but it is clear that data repositories and providers still struggle with the details of implementation. Indeed, there is still much misunderstanding about the principles, with some providers claiming FAIRness while unaware of the details of the fifteen principles underlying the broad FAIR concept.

NASA has long been a pioneer in open science and has recently published a forward looking scientific information policy (SPD-41a) that says that NASA data should be FAIR. NASA recognizes that there is nothing intrinsic to the FAIR principles that requires openness, but we do recognize that to achieve the inclusive, collaborative, and ultimately transformative aspects of open science, data must be FAIR.

To this end NASA has been working through an extensive process involving surveys, use cases, community webinars, and a dedicated workshop to define what FAIR means for NASA data. This is a very large scale endeavor given NASA's very large and bespoke data holdings that range from genomics to the origins of the universe (including petabytes of streaming Earth observation data).

This presentation reviews NASA's process and our ultimate conclusions and recommendations on how to improve understanding and alignment in the interpretation of the specific principles, what governance processes are needed, and shared guidelines for how NASA's 35 repositories can make their data programmatically FAIR, in a way that satisfies users and handles ever growing volumes and diversity.