The ARDC encourages the adoption of the FAIR principles as a valuable way of making research outputs more reusable, both for humans and machines.
The Research Data Alliance (with global input) has published a model on how to assess FAIR. This guideline translates these into a practical set of actions for ARDC projects, that will make the data more FAIR. We consider these realistic aims based on the state of maturity of services and standards currently available. The links provide guidance on how the different aspects can be achieved. Please consult with your ARDC liaison for the project as they will be able to provide guidance, support and organise training on all these aspects.
If you believe there is a good reason why a certain aspect cannot be addressed, please discuss this with your project liaison as early on as possible. The reason for it not being addressed will need to be documented and provided in the progress and final report.
Download the full guideline here [PDF, 242KB]
Findable
| Action | Required/ Recommended | Links | |
|---|---|---|---|
| 1 | Assign all data outputs appropriate PIDs (DOIs are the default identifier for this purpose) | Required | PIDs guidelines |
| 2 | The discovery metadata schema and fields to be completed for all data outputs should be discussed by partners and ARDC before data collections are created. Upon creation, ensure the agreed discovery fields are filled in in the metadata. | Required | See Appendix 1 for required fields List of mandatory and optional collection elements for RDA What makes a good collection record |
| 3 | All data outputs generated by the project must have a record in Research Data Australia | Required | Providing metadata records to Research Data Australia Defining a data collection |
| 4 | Records for all data outputs are registered with relevant discipline-specific discovery aggregators (if one exists) | Recommended | For a list of potential discovery mechanisms see Re3Data |
| 5 | The persistent identifier for the data being described must be included in the metadata | Required | For example, providing the dataset DOI in the identifier element in RIF-CS |
Accessible
| Action | Required/ Recommended | Links | |
|---|---|---|---|
| 6 | All data outputs are made as openly available as possible, they are only closed where necessary | Required | For guidance on what Open means see e.g. Open Data Handbook Expectations on Accessibility |
| 7 | All data outputs are made available through a repository | Required | See guidance on How to select a repository |
| 8 | All data outputs are available as a download and/or accessible through an open, documented API | Required where data is not closed | See guidance on APIs |
| 9 | If the data outputs are not openly available there is a clear description on the landing page on how to request access to the data outputs and conditions that need to be met | Required where relevant | See Guidance on Access Rights |
| 10 | The persistent identifier for the data output will point to a landing page about the data output, even if the data output is not public. | Required | See Best Practice for landing pages |
| 11 | If the data output is not openly available there is an authorisation and authentication procedure to provide access to the data. | Required where relevant | |
| 12 | The persistent identifier for the data output will continue to point to a landing page, even if the data output is no longer available. There is a policy to maintain these landing pages. | Required | See Best Practices for Tombstone pages |
Interoperable
| Action | Required/ Recommended | Links | |
|---|---|---|---|
| 13 | Data outputs use community-agreed standard data formats. | Required where research community agreed standard formats exist | For standards in specific disciplines see FAIRsharing.org |
| 14 | Metadata uses community-agreed standards | Required where research community agreed standard formats exist | ARDC Metadata guide |
| 15 | Data and metadata use community-agreed vocabularies, data models and ontologies (preferably internationally agreed ones where they exist) | Recommended | Good practice in vocabulary creation Research Vocabularies Australia Vocabularies and research data |
| 16 | Metadata contains persistent identifiers for research objects and entities (people, organisations) linked to the dataset(s). This includes links to ORCIDs, grantIDs, RAIDs, DOIs, IGSNs of related materials | Required | Which PIDs to use: ARDC Persistent Identifiers (PIDs): Guidelines for ARDC Investments document ARDC Citation and Identifiers guide |
Reusable
| Action | Required/ Recommended | Links | |
|---|---|---|---|
| 17 | All data outputs are assigned a machine readable (default choice should be a CC-BY 4.0) licence. | Required | ARDC Research Data Rights Management Guide |
| 18 | The licence information is available in a machine readable form on the landing page that the persistent identifier refers to. | Required | Guidance on Machine readable reference to CC licence |
| 19 | The landing page contains a citation statement for the data output | Required | ARDC Data Citation guide DataCite DOI Display Guideline |
| 20 | Provenance information on the data output is attached alongside the data | Recommended | ARDC guide on data provenance |
| 21 | Relevant discipline-specific metadata to enable reuse is captured and presented alongside the data output following research community best practice. | Recommended | See links in ARDC Metadata guide |
Download the policy here [PDF, 242KB]
