The FAIR principles of data access

An open data licence is not enough to guarantee that open data will be useful. To meet our users' expectations of the data we publish, the data must be fit for purpose

The FAIR data principles are a set of four guiding principles that help organisations make sure that the data they publish openly meets the expected quality standards. Under the FAIR principles, data must be Findable, Accessible, Interoperable and Reusable.

There are several sources that describe what each of the FAIR principles means, but one of the more accessible examples that we’ve found comes from the UK’s Geospatial Commission, in their assessment of the UK’s national geospatial data assets (figure 4):

• Findable: Findability is simply the capability of something to be found. In recent years this means data discovery through a web search engine. Data should be easily and repeatedly discovered by experts and non-experts. Having found data, it should be easy to determine whether the data is appropriate for their intended use. To support this any data should always have good discovery metadata and be published somewhere that it can be easily searched.

• Accessible: Accessibility refers to reducing the hurdles between discovering and obtaining the data. To be accessible, the terms for using the data should be clear and simple, with registration and authentication to data services minimised. Data should be designed with accessibility in mind. Wherever possible data should be turned into products that are tailored to common usage and delivery methods, using open and widely-used standards and formats.

• Interoperable: Interoperability is the ability of different geospatial systems to accurately and unambiguously exchange data. Data needs to be designed to support interoperability, which means some consideration must be given to potential systems that may want to consume the data that is published. To enable interoperability, geospatial data should be available in multiple standardised open formats. These formats should contain information to support the transformation of the data between common datums, projections and systems. Data should also contain, where appropriate, persistent, globally unique identifiers (for example Unique Property Reference Number (UPRN)) at the record or feature level.

• Reusable: Reusable is the ability for value to be derived from using the data. In practical terms this means supplying information that supports the Findable, Accessible and Interoperability aspects of the data. Usability information is typically documented as part of the dataset metadata. Metadata should be written such that both a specialist and non-specialist user can understand whether the data is suitable for their purpose. This includes clear advice on whether there are any constraints on the data’s suitability for common applications outside its primary intended use and what appropriate mitigations may be applied.