FAIR data

The FAIR principles for research data management illustrate best practice for sharing open data. They were first set out in the 2016 paper The FAIR Guiding Principles for scientific data management and stewardship.

Findable

Findable data is easily discoverable, either from a citation in a research paper or via an online search. Typically this is achieved by depositing in a data repository such as the University of Manchester’s Figshare.

Repositories assign a 'persistent identifier' such as a DOI (digital object identifier) to datasets. This provides a unique link which will be maintained into the future by repository administrators. The persistent identifier should be used whenever a dataset is cited.

Accessible

Storage in a suitable repository ensures long term preservation for your dataset. In principal Metadata should be accessible even if a dataset itself needs to be closed to the public. Even when a dataset is restricted and contains sensitive information, metadata can often be made open. The best metadata contains rich detail and is machine readable.

Interoperable

Interoperability means easy integration with other data sources. The goal is to create a “language” shared globally between different data sources. To achieve this data is given unambiguous and machine-readable contextual description.

Reusable

Reusability is at the core of data sharing, ensuring that your data is as useful as possible for any potential users. Clear licensing is key for any potential reuse of your data. This should be a permissive license by default, such as CC-BY.

Further resources

This web page is the first formal publication of the FAIR Principles, and includes the rationale behind them, and some exemplar implementations in the community.