FAIR Metrics co-creation
This page contains information on the metrics that our team has co-authored and whose homepages we host here. We encourage the re-use of these metrics whenever they are applicable to other communities. For a full list of co-authored metrics, take a look at the metrics in FAIRassist that have the Oxford e-Research Centre as a related organisation.
FAIR Metric – F2 - Metadata – discovery-oriented metadata fields
https://doi.org/10.25504/FAIRsharing.82c497
This metric evaluates whether metadata includes key descriptive elements that support discovery. F2 expects an identifier and evaluates the metadata the identifier points to for the presence of a set of core/generic fields that aid discovery. Note that F2 is concerned with discoverability only, as opposed to domain-specific completeness or reusability (which are addressed by other FAIR Principles, e.g. R1.3). This metric checks for the presence of essential discovery-enabling fields, specifically title, contributor names, summary/abstract/description, and free text tags. Although knowing what fields major search engines use to drive discovery is challenging, these metadata elements are likely to be key to discovery and are often ignored by metrics for other FAIR principles.
F3 – Metadata - uses common formats to reference data identifier(s)
https://doi.org/10.25504/FAIRsharing.0a2061
F3 requires that metadata include the identifier of the research object it describes. To align with FAIR, if a metadata record is separated from its associated research object (which is common in harvesting and indexing), the connection between them must be explicitly preserved. This metric is concerned with the explicit linkage aspect of FAIR; F3 ensures that a machine-actionable relationship exists, allowing harvesters to unambiguously navigate from a descriptive metadata record back to the research object itself. Because many research objects cannot be arbitrarily extended to include references to their metadata, in many cases the only means to discover the metadata related to a research object will be to search based on the identifier of the research object itself.
While Principle F1 focuses on the qualities of the identifier schema being used (globally unique, persistent and resolvable), Principle F3 relates to the linkage aspects of the research object identifier. It does not mandate that the identifier of the research object within the metadata record be immediately resolvable; rather, it requires that the identifier is present, and formally structured so that the relationship can be maintained and indexed by automated services.
The discovery of a research object should be possible from its metadata. For this to happen, the metadata must explicitly contain the identifier for the digital resource it describes, and this should be present in the form of a qualified reference. This metric evaluates whether the identifier provided as input resolves to metadata that contains a research object identifier. Further, that research object identifier must be distinguished from the numerous other fields and values that will be present in the metadata via one of the following formats: schema.org JSON-LD format, DataCite XML, DC XML.
FAIR Metric – A1.1 – Metadata - HTTP(S) retrieval protocol
https://doi.org/10.25504/FAIRsharing.79aee0
Principle A1.1 requires that metadata be retrievable by an open, free, and universally implementable protocol. The intent is to ensure that access to metadata is not gated by proprietary technologies, non-standardized interfaces, or restrictive licensing that prevents automated integration and universal access. Protocols such as HTTP(S) are the gold standard for this requirement because they are openly specified web standards, free for any developer to implement, and universally supported across virtually all modern software environments and programming languages. Protocols that are closed source or that have royalties associated with them could prevent users from being able to obtain the resource. By mandating these standards, we ensure that metadata remains accessible to the widest possible range of humans, harvesters and research discovery services.
This metric evaluates whether the protocol used to retrieve metadata referenced by the provided identifier is either HTTP or HTTPS and therefore openly specified, free to implement, and universally implementable.
FAIR Metric – I1 – Metadata - Database-level knowledge representation languages (syntactic)
https://doi.org/10.25504/FAIRsharing.5f33dd
What is being measured?
This metric evaluates whether there is a declared relationship (related to, implements or outputs) to a model/format from the hosting database for the provided identifier in the FAIRsharing registry, and that the associated model/format has a machine-readable structure.
Why should we measure it?
I1 requires that metadata use a formal, accessible, shared and broadly applicable language for knowledge representation. The purpose of this principle is to ensure that metadata is expressed in a standardised format that can be interpreted and exchanged across systems. In this metric, the definition of syntactic and semantic formats follows the FAIRsharing guidance. This metric evaluates whether there is a declared metadata model or format that ultimately resolves to a basic representation format that is considered syntactic by the FAIRsharing registry. Note that all formats that match the guidance on semantic formats also fulfil the requirements for being syntactic formats.
Communities may wish to use this metric in conjunction with others that examine the precise format of the resolved metadata.
Note that, in order for this metric to function, tests must know the FAIRsharing database record URL/DOI so that FAIRsharing metadata may be retrieved. This can be done in a variety of ways, e.g. using the FAIRsharing API to find possible matches between the identifier and registered databases, hard coding the FAIRsharing record identifier, or expecting it to be provided as a second argument to the test.
FAIR Metric – I1 – Metadata - Database-level knowledge representation languages (semantic)
https://doi.org/10.25504/FAIRsharing.9114a7
What is being measured?
This metric evaluates whether there is a declared metadata model or format that is based in a representation format that is considered grounded / linked data within the FAIRsharing registry (for more information see the FAIRsharing documentation).
Why should we measure it?
I1 requires that metadata use a formal, accessible, shared and broadly applicable language for knowledge representation. The purpose of this principle is to ensure that metadata is represented in a widely recognised and interoperable format. In this metric, the definition of syntactic and semantic formats follows the FAIRsharing guidance. This metric evaluates whether there is a declared metadata model or format that ultimately resolves to a basic representation format that is considered grounded / linked data by the FAIRsharing registry.
Communities may wish to use this metric in conjunction with others that examine the precise format of the resolved metadata.
Note that, in order for this metric to function, tests must know the FAIRsharing database record URL/DOI so that FAIRsharing metadata may be retrieved. This can be done in a variety of ways, e.g. using the FAIRsharing API to find possible matches between the identifier and registered databases, hard coding the FAIRsharing record identifier, or expecting it to be provided as a second argument to the test.
FAIR Metric – I2 – Metadata – has FAST subject heading
What is being measured?
This metric evaluates whether there is at least one FAST subject vocabulary term used in the metadata that resolves to content in a linked data format (see FAIRsharing guidance).
Why should we measure it?
I2 requires that metadata use vocabularies that follow FAIR principles. The purpose of this principle is to ensure that descriptive terms are drawn from such vocabularies, enabling consistent interpretation across systems and communities. Under this semantic interpretation, the evaluation checks whether there is at least one FAST subject vocabulary term used in the metadata that resolves to content expressed in a linked data format. The resolved term (for example, http://id.worldcat.org/fast/819673) must return structured content in a recognised linked data representation, with concepts and relationships defined. A resource satisfies this metric if at least one metadata element references a FAST subject vocabulary term that resolves to content in a linked data format (the definition of semantic formats follows FAIRsharing guidance).
FAIR Metric – I2 – Metadata – has ISO language
What is being measured?
This metric evaluates whether there is at least one ISO 639-2 / 639:2023 language code term used in the metadata that resolves to content in a linked data format (see FAIRsharing guidance).
Why should we measure it?
I2 requires that metadata use vocabularies that follow FAIR principles. The purpose of this principle is to ensure that descriptive terms are drawn from such vocabularies, enabling consistent interpretation across systems and communities. Under this semantic interpretation, the evaluation checks whether there is at least one ISO 639-2 / 639:2023 language code term used in the metadata that resolves to content expressed in a linked data format. The resolved term (for example, http://id.loc.gov/vocabulary/iso639-2/eng) must return structured content in a recognised linked data representation, with concepts and relationships defined. A resource satisfies this metric if at least one metadata element references an ISO 639-2 / 639:2023 language code term that resolves to content in a linked data format (the definition of semantic formats follows FAIRsharing guidance).
FAIR Metric – R1.2 – Metadata - creator ORCiD
What is being measured?
This metric evaluates whether there is at least one dc:creator with an ORCID in the metadata retrieved upon resolution of the provided identifier, providing one facet of evaluation for this R1.2.
Why should we measure it?
R1.2 requires that metadata be associated with detailed provenance. The purpose of this principle is to ensure that the origins and responsible individuals associated with a digital object are clearly identified, supporting attribution and transparency. This metric evaluates whether there is at least one dc:creator with an ORCID in the metadata retrieved upon resolution of the provided identifier, providing one facet of evaluation for this R1.2.
Under this interpretation, the evaluation checks whether the resolved metadata includes an ORCID identifier associated with a person in the defined dc:creator role.
FAIR Metric – I3 – Metadata – Qualified References to Related Research Objects
FAIR Metric – I3 – Metadata – Qualified References to Versioned Records
Last updated
Was this helpful?