Aller au contenu principal

Research data management

Recommandations

  • Use unique perennial identifiers such as DOI, ORCID
  • Describe data using rich metadata, controlled vocabulary and recognized standards
  • Choose open, non-proprietary file formats
  • Define data access conditions
  • Deposit data in an open data repository
  • Assign a copyright license that allows reuse

The FAIR principles define characteristics that make data and metadata easy to find, accessible, interoperable and reusable. FAIR data is not necessarily synonymous with open data. Some data, although deposited as open access, may not conform to the FAIR principles. Conversely, some restricted or closed-access data may comply with the FAIR principles.

Findable

To make data and metadata easy to find by humans and computer systems4-7 we need to :

  • Attribute a unique perennial identifier (PID) (e.g. DOI, ORCID...)
  • Describe data using rich metadata that complies with recognized standards, document the context of creation, conditions of sharing (license) and reuse, to enable proper interpretation. This information must be readable by the IT system
  • Clearly  mention  the  unique  perennial  identifier  in  the  metadata,  so  that  the association between the metadata file and the data set is explicit
  • Save data and metadata in a repository that indexes and searches metadata

Accessible

To facilitate access, consultation and the downloading of data and metadata by humans and computer systems, it is necessary to store them permanently. 1-4-7 Accessible does not necessarily mean open or free, but only that the exact conditions of access are clearly stated.4 To ensure that data is accessible :

  • Choose a data repository that uses a standardized, free and open communication protocol to facilitate data retrieval (e.g. http, ftp, smtp...)
  • Define access conditions: license, right of reuse
  • Make metadata accessible even if the data cannot be or is no longer available. (e.g. metadata on authors, institutions and associated publications can be useful even if the data is missing). This ensures data discovery and helps limit storage costs
  • Ensure that metadata and data are physically separate files to facilitate continuity of metadata discovery in the event of data deletion
  • Create an account on the selected repository to be authenticated as the author, and leave the possibility of being contacted for further information or authorization of use if data access has to be conditional 

Interoperable

To make data and metadata combinable with each other, usable and interpretable by different IT systems - we speak of technical and semantic interoperability2 - it is necessary to: 

Reusable

To make data and metadata reusable for future research, they need to be sufficiently described, in particular with regards to data provenance and conditions of use 4-7. Care must therefore be taken to :

  • Describe both the dataset (content of the dataset, how it is generated, processed and reused, etc.) and the data (information on how to use the data, e.g. definition of variable names, etc.)4-7: 
    • Describe the scope of your data: for what purpose was it generated/collected?
    • Mention any particularities or limitations of the data that other users should be aware of
    • Specify date of data set generation/collection, laboratory conditions, who prepared the data, parameter settings, name and version of software used
    • Specify if the data is raw or processed
    • Ensure that all variable names are explained or self-explanatory (i.e. defined in the controlled vocabulary of the research domain)
    • The version of archived and/or reused data is clearly specified and documented.
  • Describe the license for reusing the data in the metadata. Suggested licenses are CC BY, CC 0... 7

References

  1. Centre pour la communication scientifique directe. (s. d.). Principes FAIR. CCSD. https://www.ccsd.cnrs.fr/principes-fair/
  2. Ceris-Institut Pasteur. (2022, juin 3). Comment rendre ses données interopérables ? Open science : évolutions, enjeux et pratiques. https://openscience.pasteur.fr/2022/06/03/comment-rendre-ses-donnees-interoperables/
  3. Deng, S. (s. d.). UCF research guides : metadata: dataset metadata checklist. https://guides.ucf.edu/metadata/datasetmetadata_checklist
  4. GO FAIR. (s. d.). FAIR Principles. GO FAIR. https://www.go-fair.org/fair-principles/
  5. Observatoire global du Saint-Laurent. (2019, octobre 17). Principes FAIR. OGSL. https://ogsl.ca/fr/principes-fair/
  6. Paquette-Bigras, E. (s. d.). Documentation des données. Bibliothèques - Université de Montréal. https://bib.umontreal.ca/gerer-diffuser/gestion-donnees-recherche?tab=5243848
  7. Swiss National Science Foundation. (s. d.). Explanation of the FAIR data principles. https://www.snf.ch/SiteCollectionDocuments/FAIR_principles_translation_SNSF_logo.pdf
  8. Wilkinson, M. D., Dumontier, M., Aalbersberg, Ij. J., Appleton, G., Axton, M., Baak, A., Blomberg, N., Boiten, J.-W., da Silva Santos, L. B., Bourne, P. E., Bouwman, J., Brookes, A. J., Clark, T., Crosas, M., Dillo, I., Dumon, O., Edmunds, S., Evelo, C. T., Finkers, R., … Mons, B. (2016). The FAIR Guiding Principles for scientific data management and stewardship. Scientific Data, 3, 160018. https://doi.org/10.1038/sdata.2016.18