Search dataset descriptions

[ddeboer]

• SPARQL based on #6.

Questions

• @coret Do you have any example queries? How will users be searching for datasets?
• @coret Should this return the descriptions in DCAT or should we preserve the original dataset description’s ontology?

[coret]

Answers (also added to design document):

• (Functional) example queries

1. Search for datasets which have a SPARQL endpoint as distribution
2. Search for SPARQL endpoint to be used in terms-network (how??)
3. Search for datasets created/published/updated within last month
4. Search for datasets which have “Bevolkingsregisters” in their description
5. Search for datasets which have “wo2” as genre
6. Search for datasets published by the “Open Archieven” organization TODO: rewrite to SPARQL

• How will users be searching for datasets?

Main users are machines, like terms-network. The demonstrator can link to some (stored) example queries in TriplyDB for human consumption.

• Should this return the descriptions in DCAT or should we preserve the original dataset description’s ontology?

The 'Collected dataset descriptions' are based on the DCAT vocabulary. Crawled datasets are converted tot DCAT if schema.org (or other in the future), orginal is not stored. The used vocabulary is noted in the 'Registered URI’s of dataset descriptions' for statistical purposes.

[ddeboer]

Searching dataset descriptions is possible at https://triplestore.netwerkdigitaalerfgoed.nl.

@coret Can we refer interested parties to that or does the Registry itself need to offer search capabilities?

[ddeboer]

Discussed this with Bob, for now we will wait whether people actually need separate search functionality.

• We will refer more advanced users/researchers to the SPARQL endpoint.
• Bob has created some visualisations in the Registry demo, which we can show to other interested parties to demonstrate that we can do analysis and draw conclusions from the submitted dataset descriptions.

[EnnoMeijers]

I suggest to use the datastory approach for further presentations of the registry data (and replace the old registry datastory https://github.com/netwerk-digitaal-erfgoed/stories/tree/master/html/hackalod/register) until we have a clearer view on a possible UI (if any)

[coret]

Will investigate the datastory approach, do not known if the YASGUI component works with GraphDB. First step is enabling CORS.

[coret]

• The demonstrator has a Data Stories page. This needs to be expanded.
• The triplestore has a "saved queries" section which is shown on the startpage of the triplestore.

[ddeboer]

Can we consider this done for now?