How to make your data FAIR

This page will show you how you can make your research data more FAIR by taking you through six FAIRification practices 

  1. Documentation 
  2. File formats 
  3. Metadata 
  4. Access to data 
  5. Persistent identifiers 
  6. Data licences 
image description

Case introductions

Throughout the FAIRification subpages, four research projects are used as examples of how you can make your research data more FAIR. You will see short interview clips of researchers in Engineering, Humanities, Health Sciences, and Social Sciences. The researchers share their methods and solutions for problems specific to quantitative data, qualitative data, and sensitive data. Watch all four case introductions - or just the ones you expect will be most relevant to your research field and data type.  

Quantitative data

Nikola Vasiljević works with unique measurement data of the wind field behind wind turbines, or wake. Having FAIR data is essential to his field, because this kind of data has a high potential to be reused for many decades, from the very first day they are created.

Qualitative data

The Language Technology Group works with transcript data from the Danish Parliament. They publish their data, the Danish Parliament Corpus (2009-2017), in a FAIR way in CLARIN-DKHaving FAIR data is essential to their field, because without freely accessible data they could not perform their research. 

Sensitive quantitative data

Carsten Brink works with sensitive personal data in the field of radiotherapy for cancer patients. Having FAIR data is essential to the researchers in his field, because they need interoperable data from a large number of patients to predict outcomes. The researchers in his field used to gather data from many different institutes and pool them in one physical location to run models. But because they need to preserve the patients’ confidentially, physically moving the data is a legally complicated task. Instead, they now work with distributed learning, which allows them to analyse data at other institutes without having to physically move them.

Sensitive qualitative data

Ditte Shamshiri-Petersen works with sensitive survey data in an international collaboration called the International Social Survey Programme, or ISSP. Having FAIR data is essential to her field, because without findable documentation and metadata of the surveys conducted in each member country, they could not perform cross-country longitudinal studies.

FAIRification practices

We recommend you start with "Documentation" and work your way through all six of the FAIRification practices from left to right, top to bottom, one by one.

In summary...

image description

Findable

You make your research data findable for your collaborators and the rest of the world by: 

  • Publishing your data and/or metadata in a searchable resource such as a repository like Dataverse, Zenodo or Figshare that assigns a persistent identifier

  • Including rich accurate machine-readable descriptive metadata and keywords to your data, preferably according to a community-specific metadata standard (e.g. Dublin Core) or ontology
image description

Accessible

You make your research data accessible by: 

  • Attaching a data licence or clear data accessibility statement in your openly available administrative metadata 

  • Ensuring your data are archived in long-term storage and retrievable by their persistent identifier using a standard protocol

  • Giving access to the metadata, even if the data are closed
image description

Interoperable

You make your research data interoperable by: 

  • Including sufficient and standardised structural metadata in accordance with your research community’s standard controlled vocabulary or ontology 

  • Including use of common standards, terminologies, vocabularies, ontologies and taxonomies for the data 

  • Preferring open, long-term viable file formats for your data and metadata 

image description

Reusable

You make your research data reusable by: 

  • Attaching all the relevant contextual information required for re-use in either the documentation or metadata attached to your data 

  • Including sufficient and standardised structural metadata in accordance with your research community’s standard controlled vocabulary or ontology 

  • Preferring open, long-term viable file formats for your data and metadata

  • Applying a machine-readable data licence