Hey researchers, is your data FAIR?
FAIR, as in per the Cambridge Dictionary's definition of fair: "Treating someone in a way that is right or reasonable, or treating a group of people equally and not allowing personal opinions to influence your judgment?"
Um, no...but come on, you know you love a good acronym - and this IS a good acronym 😉
FAIR stands for Findable, Accessible, Interoperable and Reusable. It's all about maximising the benefits data collected can give - making sure we give it (and you) maximum impact.
That's not to say that data has to be fully open. There are times when data - or some parts of a dataset at least - need to be a bit more privacy-orientated. Think personal health data, locations of highly endangered species that are targeted by poachers, and the like. Let's break down what each component of FAIR is:
🔎 Findable
Can your data actually be found by others, or do the only people who know of its existence reside within your lab...or just your brain?
🔓 Accessible
People know your data is out there, but can they actually access it, or is it stuck on a hard drive somewhere in one of those boxes in the corner of that dusty cupboard that no longer works?
📖 Interoperable
Did you put your data in a recognised standard (formats, languages, ontologies, etc.), allowing it to be used with outer datasets, or did you decide to make up something else completely different because you fancied being, well, fancy?
♻️ Reusable
Can others actually use the data to do cool, amazing science and guide decision-making and the like, or have you decided they can look but not touch?
The FAIR data principles. Credit Patrick Hochstenbach/Open Science Training Handbook (CC0 1.0 Universal)
Starting FAIR is easier than trying to be FAIR later
It might sound like a lot of work, but if you think about making your data FAIR at the beginning, it really isn't. Much of making data FAIR means having things like decent metadata, storing your data in a data repository that isn't stuck behind a paywall, persistent identifiers (a bit like the DOI on your peer-reviewed papers), and clear use licences. For data with elements that need to be kept private for whatever reason, there are options to do that and still offer some FAIR data (this is where a good data management plan can come in really handy).
Curious to know more?
Here's a couple of resources:
👉 "FAIR data. What is FAIR and why is it important?" FAIR data — Ghent University (ugent.be)
👉 The Go FAIR website https://www.go-fair.org/