Test request - Dataset is cited in publication

[markwilkinson]

Test request - Dataset is cited in publication

Can this be considered as a FAIR test ?

Test

• input :
  • dataset DOI
• output :
  • if dataset is cited in at least one publication returns ""Test success publication citing this dataset was found"" else ""No publication citing this dataset was found

Implementation

3 resources should be used

• OpenAlex Citation in full text (typically in data availability section)

  • https://api.openalex.org/works?filter=fulltext.search:%2210.15151/ESRF-DC-572252655%22 (1 pub in this example)
  • limitation, OpenAlex does not index supplement materials docs, DOI may be cited there)

• OpenAlex Citation in references  section

  • DataCite datasets are indexed in OpenAlex and linked to pubs since 2025/11
  • Step 1 : get the OpenAlex id from the DOI : https://api.openalex.org/works/doi:10.15151/ESRF-DC-572252655 => https://openalex.org/W6905148043
  • Step 2 : query cites endpoint https://api.openalex.org/works?filter=cites:W6905148043 (3 pubs cited in this example)

• OpenCitations

  • OpenCitations is focused on pub to pub citation but may have a few pub to dataset links.
  • Success : https://opencitations.net/index/api/v2/citations/doi:10.15151/esrf-dc-572232527
  • Fail : https://api.opencitations.net/index/v2/citations/doi:10.15151/ESRF-DC-572252655

Metric

TODO in FAIRsharing

[rduyme]

FAIRsharing metric : https://fairsharing.org/7466

[markwilkinson]

Definitely not a test of FAIRness ("popularity" has long been excluded from things that FAIR cares about).

However, it can be a Community metric, if that's important to your community. It feels quite... unfair... since it is entirely out of the owners control... but we can build the test if you want it! But if you fail... what will the advice be? "Do more advertising"?

I would prefer to keep this one low-priority, if that's OK

[rduyme]

In instrument facilities we generate datasets, based on selected scientific project proposals. The main purpose of these datasets is to produce science often via a scientific paper. We expect the scientists to produce analysis. The primary paper written by the experimenter adds context and information on how this dataset was initially created, analysed etc...
This initial paper is quite central for the future reuse of the dataset by other scientists. ("R" of FAIR )

The common way to link a dataset is to use the "Data Availability" section
https://doi.org/10.1016/j.foodhyd.2025.112037

A random dataset without any reference publication analysis, is probably often not very reusable. We could even think of deleting its content, it in the future.

Note : In astronomy, marine observatories you may not have this initial expectation, you may produce raw sky or ocean data without initial project in mind.

[markwilkinson]

If you create the Metric, I will write the test!

Note that I have FINALLY found the time to create the DCAT Turtle versions of the core FAIR Metrics, and also the metrics that you have created for ESRF:

https://github.com/FAIRMetrics/Metrics/tree/master/docs/metrics

If you want to spend time making a new Turtle for your metric (in the esrf subfolder) at the same time as you register it in FAIRSharing, that saves me time/effort. If not, please just go ahead and register with FAIRsharing and I will copy over the information into a turtle file.

[rduyme]

FAIRsharing metric : https://fairsharing.org/7466 exists.

Do we need to add a link to this new github ttl in the fairsharing record as well ?

[markwilkinson]

Please raise this at the meeting tomorrow - this is EXACTLY the kind of question that we don't yet have a clear answer for!