You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Anyway, I think the issue name is misleading, JSON-LD is the format, but schema.org/Dataset (+ Google extensions?) is the actual schema. It seems that Google will start using it to discover datasets from 3rd parties, so that alone might signal future adoption with high probability, and also schema.org stuff usually ends up being more used in the long term.
With respect to the format itself, we might prefer JSON (afaik JSON-LD is valid JSON) for convenient parsing of metadata rather than XML.
Note that I have no strong preference for JSON-LD itself, since I never really used it. But I have a preference for adopting schema.org et al vocabularies as well as JSON over XML.
So, Where is Your Dataset?
It is probably clear by now that Dataset Search is only as good as the metadata that exists on the Web pages for datasets.
The most common answer to the question of why a specific dataset does not show up in our results is that the Web page for that dataset does not have any markup. Just pop that page into the Structured Data Testing Tool and you will see whether the markup is there. If you don't see any markup there, and you own the page, you can add it
Activity
smola commentedon Apr 26, 2018
@campoy has been working on a proposal to add metadata to our datasets: src-d/guide#163
it would be good to add this info to that discussion.
campoy commentedon May 4, 2018
I'm curious, what are the benefits of JSON-LD over other formats such as PMML?
smola commentedon May 4, 2018
@campoy Is PMML used for dataset metadata at all?
Anyway, I think the issue name is misleading, JSON-LD is the format, but schema.org/Dataset (+ Google extensions?) is the actual schema. It seems that Google will start using it to discover datasets from 3rd parties, so that alone might signal future adoption with high probability, and also
schema.orgstuff usually ends up being more used in the long term.With respect to the format itself, we might prefer JSON (afaik JSON-LD is valid JSON) for convenient parsing of metadata rather than XML.
campoy commentedon May 16, 2018
I don't have much experience on this, so if @smola has a preference for JSON-LD and Google is also using it, I say let's go with that.
smola commentedon May 18, 2018
Note that I have no strong preference for JSON-LD itself, since I never really used it. But I have a preference for adopting
schema.orget al vocabularies as well as JSON over XML.dennwc commentedon Sep 30, 2018
Related: https://ai.googleblog.com/2018/09/building-google-dataset-search-and.html
bzz commentedon Oct 25, 2018
Yes, basically if we could just annotate dataset homepage with structured information https://search.google.com/structured-data/testing-tool and it have a good chances of being indexed.