Best practice for publishing classifications as linked data with XKOS
Franck Cotton  1@  
Institut national de la statistique et des études économiques (INSEE)

XKOS is an RDF vocabulary extending SKOS and dedicated to the representation of statistical classifications. Though XKOS recently celebrated its tenth anniversary, it was officially published in 2019 by the DDI Alliance and the XKOS Working Group was formally created in 2021.

Also in 2021, the production of a "Best Practice" document was started. This guide intends to provide advice on how to use XKOS for maximum interoperability and reusability, and will cover topics like the description of classifications (labels, explanatory notes, levels), the different types of correspondences, how to represent evolution over time of the different elements, etc.

A number of specific topics will also be addressed, for example how to tackle multiple languages, how to publish statistical classifications as XKOS, or what kind of descriptive metadata to attach in order to maximize findability.

Where appropriate, best practice rules will be formalized in the SHACL validation language, so that publishers can check if their classifications conform to the guidance. In addition, the guide will propose a list of useful tools for the production and publication of XKOS statistical classifications.

The presentation will provide an overview of the XKOS specification, then detail the progress made regarding the best practice guide.

Online user: 2 Privacy