In 2014, members of the scientific community got together in a workshop called “Jointly Designing a Data Fairport” in Leiden, the Netherlands, to define the principles on how you could make your data more findable, accessible, interoperable, and reusable.These 15 principles are now called The FAIR Principles. The GO-FAIR organization has noted the principles and their meaning.
The idea behind these principles is that the data itself should not be used just once, but (if possible) many times. This gives more value to the data itself, and the time it took to produce these data.
Following the FAIR Principles, it makes sure that your data can be found, both by humans and by machines. Additionally, it also makes sure that the machine (or human) could understand who and under what conditions can use the data (both in legal aspects and in a practical sense) and how the data could be used. What must be noted is that none of the FAIR Principles are mandatory, and you can apply them in any research domain.
Another important aspect is not to confuse FAIR data with Open data. Open data (and science) idea is that all the data (or science resources in general) should be freely available for everybody to use in any capacity (this means no restrictions). However, researchers cannot always do that. For example, sensitive information in many cases cannot be fully open. This is where the FAIR Principles shine. It is understood that not all the data can be fully accessible for everybody, but they still can be FAIR. For example, the metadata (information about the dataset) can (and should) be findable and accessible, even when the dataset is not available, or the dataset has been deleted.