What are the FAIR principles? 

The FAIR principles, introduced in 2014 and formally published in 2016, serve as a framework to meet the challenges posed by the increasing reliance on computational support for data analysis and interpretation. These principles emphasise the importance of making data findable, accessible, interoperable, and reusable by both humans and machines. By adhering to FAIR principles, researchers and institutions can ensure that their data are well-described, easily discoverable, and ready for reuse in diverse contexts. 

How can I make my data FAIR? 

Findable  

Findability refers to the ease with which data can be discovered and located 

Practical steps: 

  • Deposit your data with a trusted data repository. This will ensure that your dataset is assigned a unique persistent identifier (e.g. DOI). Depositing your dataset with a trusted repository will also ensure that the metadata for the dataset is indexed in a searchable resource (e.g. Google Search) 
  • Describe your data with ‘rich metadata’. This can be included in the published metadata record or added separately as a README text file.  
  • Suggest a recommended citation for your dataset in the metadata, and include the persistent identifier when citing the dataset 
  • Include a data access statement in associated publications giving details of how and where the dataset can be accessed and any restrictions on accessing the data 

Accessible 

Accessibility ensures that data can be retrieved and accessed by both humans and machines 

Practical steps: 

  • Data should be retrievable by their identifier using a standardised communications protocol. If you deposit your data with a trusted repository, it will have a standard protocol in place (e.g. http) to support online retrieval 
  • There may be valid reasons why data cannot be shared openly (e.g. to protect data privacy/confidentiality). If possible, choose a repository that enables controlled access to sensitive data 
  • If access to the data is restricted, publish a metadata-only record and include details of how the data can be accessed and under what conditions 
  • If the data are longer available, the metadata should still be accessible 

Interoperable  

Interoperability enables seamless integration of diverse datasets, applications, and workflows 

Practical steps: 

  • Use open or widely used file formats wherever possible. 
  • Make use of common standards for metadata, vocabularies, keywords, and ontologies if available   
  • Include clear links to articles and other related research resources in the metadata and use persistent identifiers (e.g. DOIs) where available  

Reusable  

Reusability ensures that data can be used for various research purposes beyond their initial collection 

Practical steps: 

  • Include detailed information about the provenance of the data i.e. how, why, and by whom the data were created and processed. This can be included in a README text file.  
  • Make sure the metadata includes a ‘rich’ description of the attributes of the dataset e.g. definitions and explanations of variables, codes, software dependencies.  
  • Include a data dictionary for tabular data either separately or as part of a README file.  
  • Include any supporting documentation that will help others understand and reuse the data (e.g. protocols, reports, blank consent forms, survey tools).  
  • Use discipline-specific metadata standards where available  
  • Make sure your data are available under open licenses, such as Creative Commons licenses, to maximize potential reuse 

FAIR v open 

FAIR data does not necessarily mean open access to data. Not all data can be made openly available. Sometimes restrictions are needed to protect sensitive and confidential information (see our web page Sharing sensitive data). However, data that are not openly available can still be FAIR if the guidelines outlined above are followed e.g. if the data are accompanied by metadata and supporting documentation which provide detailed information about the providence of the data and how and under what conditions the data can be accessed. The principle you should follow is for the data to be ‘as open as possible, as closed as necessary’  

Planning for FAIR data 

It is important to embed FAIRness into your data management practices from the start of your project, and creating a data management plan can help you achieve this. Implementing a data management plan helps to enhance the FAIRness of your data from the outset.