ABSTRACT

Data linking and integration are approached from the perspective of the FAIR principles: How to render data findable, accessible under welldefined conditions, including basic licensing information, interoperable and therefore reusable. Linking data in a functional way, so that they effectively serve as a single data source for distributed analytics, is a very powerful way to avoid classical, cumbersome, and error-prone 'extract, transform, load' (ETL) processes over and over again. The classical approach to 'data integration' is a very logical and in some cases still valid approach: Relevant data may be dispersed over many different databases. Workflows for FAIR data handling are increasingly available. For most data types and metadata needs, there are earlier examples, emerging community standards, and publication, as well as linking, integration, and analytics workflows that can handle FAIR data. In any case, data should be 'published' in such a way that they become optimally findable, accessible, interoperable, and reusable for others.