The FAIR principles were published in 2016 and have since become a cornerstone of EU data policy. They guide how data should be managed to maximise its value — particularly for scientific research and public sector data. Understanding FAIR is essential for EPSO data management competitions.
Findable
Data must be easy to find for both humans and machines:
- Assign a globally unique and persistent identifier (like a DOI)
- Describe data with rich metadata
- Register data in a searchable resource (catalogue, repository)
- Metadata must clearly include the identifier of the data it describes
EU context: The EU Open Data Portal and data.europa.eu implement findability through DCAT-AP metadata standards and persistent URIs.
Accessible
Once found, data must be retrievable:
- Data is retrievable by its identifier using a standardised, open protocol
- The protocol allows for authentication/authorisation where necessary
- Metadata remains accessible even when data is no longer available
Key distinction: Accessible does NOT mean open. Data can be FAIR and still require authentication. The point is that the process for access is clear and standardised.
Interoperable
Data must be able to integrate with other data and systems:
- Use a formal, accessible, shared, and broadly applicable language for knowledge representation
- Use vocabularies that follow FAIR principles themselves
- Include qualified references to other data
EU context: This connects directly to the EIF semantic layer. Core Vocabularies, controlled vocabularies, and common data models (like DCAT-AP) enable interoperability across EU systems.
Reusable
Data must be well-described so it can be replicated or combined:
- Rich metadata with clear and accessible data usage licence
- Detailed provenance information
- Data meets domain-relevant community standards
EU context: The European Commission promotes CC-BY 4.0 and CC0 licences for open data reuse. The Data Governance Act further facilitates reuse of protected public-sector data.
FAIR vs Open Data
A common EPSO exam trap: FAIR data is not necessarily open data. FAIR is about making data findable and accessible through clear mechanisms — but the data itself may require authorisation to access. Open data is always FAIR, but FAIR data is not always open.