Reproducible Research › Module 5: Data and models

FAIR data

Lesson 1 of 3 · 2 min

Previous

In 2016, a group of researchers and publishers proposed the FAIR principles for sharing data. They are now used by funders and journals worldwide.

The four letters

  • Findable: the data has a permanent identifier (like a DOI) and a description that search engines can find.
  • Accessible: people (and machines) can retrieve it by its identifier, using standard methods. If access must be restricted, the conditions are clear.
  • Interoperable: it uses common formats and vocabularies (CSV, standard units, standard codes) so it can be combined with other data.
  • Reusable: it has a clear licence, a description of where it came from, and enough detail to be used correctly.

FAIR is not the same as open

Sensitive data, like medical records, can be FAIR without being public: the description is findable, and there is a clear process to request access.

Practical steps

  1. Use open formats: CSV instead of a proprietary format; describe every column.
  2. Write a data dictionary: each variable, its meaning, units and allowed values.
  3. Put the data in a repository that gives a DOI: Zenodo, Dryad, OSF, or a field-specific one.
  4. Add a licence, for example CC BY 4.0 or CC0.

RememberFindable, Accessible, Interoperable, Reusable: a DOI, open formats, a data dictionary and a licence.