@rickmoynihan I have watched the presentation, read what's available on http://csvw.org, and read the CSVW primer document on the w3c website. I think it's really cool how the metadata is totally separate from the data, allowing you to basically annotate external CSV files too in addition to the ones you control yourself. I will definitely attempt to create metadata files for each of my CSV exports. However, while it is obviously underpinned by RDF (disguised as JSON-LD), it does seem like it may not be quite as flexible as I had hoped with regard to support for qnames—and supporting RDF resources use in general... Or at the very least this facet isn't as well-described in the documentation I've consumed so far. For example, if I have a column where the values are qnames which might use several different (known) namespaces, I see no way of reconstructing these into different RDF resource IRIs using the semantics of the metadata JSON file. :thinking_face: So I guess my only choice is representing them as fully realised IRIs in that particular column or having a separate column for each namespace, so that I can use the aboutURL string interpolation method described in the primer. I wish there were a way to provide a prefix mapping in the metadata file, so that I can just use qnames directly in my columns.

e.g. what I want is some CSV output that can work like this

dn:someResource, dns:someRelation, dn:otherResource
dn:anotherResource, wn:otherRelation, dn:yetAnotherResource