How to provide an RDF Representation of the schema of a dct:Distribution (CSV file) ?

[lorenae]

Hi!
After reviewing the discussions related to this topic (issue 1418 , issue 1426) , I have come to the conclusion that the recommended approach to associate a CSV file with its schema specification (mainly columns) is to use the dct:conformsTo property to link the dct:Distribution as a CSV of a dataset to a JSON document following the Model for Tabular Data and Metadata on the Web specification.

While this approach effectively represents the structure, it does not provide the schema in RDF itself. Ideally, I would like to create an RDF resource for each column in the CSV and link each of those resources to the concepts the columns represent.

My question is:

1. Does DCAT provide any recommendations or guidance for achieving this?
2. Are you aware of any existing RDF vocabularies that could be used for this purpose?

Thank you for your time and assistance!

Best regards,
Lorena

[pwin]

Hi Lorena
Thanks for starting this discussion - it's quite an important aspect of discussion. One approach you might use is to publish a TARQL script ( https://tarql.github.io/ ) - it's not RDF, it is using SPARQL syntax, and perhaps there is some scope for expanding the SPIN ( https://www.w3.org/submissions/spin-sparql/ ) approach to describe it in RDF.
What do you think?
Cheers
Peter