FAIR Data

ITA | ENG

According to Open Science funding programs and policies, data must have specific characteristics, must be:

• Findable: trackable and described using standard metadata sets (e.g., Dublin core) or metadata by discipline to describe specific materials clearly. They must use unique identifiers (e.g., DOI) to enable trusted automatic retrieval.
  They must be stored for at a minimum of 10 years in a repository that offers guarantees of security, persistence, and can also be indexed by machines.
• Accessible: open as often as possible, as closed as necessary), always and in any case accessible to those who make a justified request. Metadata accompanying them, or their description, should remain available at any time even when the data are not in open access.
• Interoperable: readable, reprocessable, and combinable with other data from FAIR-compliant systems. Data should have certain characteristics, such as being in non-proprietary, uncompressed, unencrypted formats with documented standards. Metadata should also use a standardized language shared globally by different indexing services.
• Reusable: provided with documentation useful to be understandable and reusable (provenance, methodology, tools or software with which they were generated, protocols, ...) and a licensing statement with permitted uses.

Sensitive data and privacy

Data are not subject to copyright, however, a particular expression of data, such as a graph or table in a book or dataset, may be.
In order to validate and promote research, it is critical to maintain the confidentiality of research subjects through the identification and proper management of potential conflicts between ethical and privacy requirements and conditions for depositing data in a repository.

Considering share and reuse data, they can be:

• Closed data: usually personal or commercially highly sensitive data that cannot be shared except in repositories where only authorized users have access.
  Closed access must be justified and applicable to specific well-defined cases (protection of privacy of persons involved in the research, security reasons, production of sensitive data, economic exploitation of research results/patents) or otherwise well-motivated.
• Shared data: can be made widely accessible but not necessarily to everyone, even only to specific groups. The license must be clearly stated to make explicit the possible uses.
• Open data: data that can be used, modified and shared by anyone, for any reason, as long as source and open access are preserved. A license must be clearly stated that makes explicit its possible uses.

To ensure its security, data should be kept in multiple copies and in secure environments even during research, not just when completed.

Tools

• Checklist EUDAT
• Checklist ARDC
• Checklist DANS

More information

• Fair Aware DANS
• FAIRsFAIR Data Object Assessment
• Foster course on Fair Data