FAIR data

FAIR is an acronym for Findable, Accessible, Interoperable, and Reusable, and the principles aim to make research data more accessible and easier to reuse. Since their launch in 2016, the FAIR principles have gained widespread traction, and many research funders refer to them in their guidelines.

The principles emphasise that research data should also be possible for digital systems to find, access, interact with, and reuse with minimal human intervention. This is because the constantly increasing volume and complexity of research data require that digital tools can be used to simplify the management.

The FAIR principles are general, and how researchers make their data FAIR can vary between different scientific disciplines and types of data. It is easier if the planning of a project already includes how the data will become FAIR.

Some general tips for making data FAIR

Findable

In order for data to be searchable and findable, they should be well described and registered in established data repositories. The dataset then receives a permanent identifier, such as a DOI number, and the description (metadata) becomes searchable and openly accessible on the internet. Use established standards and terminology within the domain when describing the data. See also Publish data.

Accessible

The FAIR principles are an important component in the transition to open science. Ideally, data should be directly available for download without embargoes or other restrictions, but a dataset can be FAIR without being openly accessible. If the data cannot be shared openly, it can still be described in a repository. Include information about the restrictions and any access conditions, as well as contact details.

Interoperable

Make it easy for others to use your data in their own digital environment. Choose open, non-proprietary standards for formats and data types whenever possible. Use established practices and standards within your field when choosing and naming variables and values.

Reusable

It should be clear under what conditions and type of license the data can be used. Well documented data is important for reusability – what information is needed for someone else to understand and reuse your data? It may include how the data has been generated and analysed, how the data is structured, as well as explanations of the variables and values used. Also inform about what other resources may be needed to review or reuse the data, such as code, scripts, protocols, and questionnaires.