RE: Dataset discovery project @ Google (was: Re: Joint session on vocabulary with Web of Things)

Andrea and Dan,

And a Met-Ocean specific filter: 100 simultaneous forecasts (or analyses, or hindcasts). We may want to choose not the most likely forecast, but the most extreme (e.g. <5% or >95% percentiles), albeit less likely to occur.

This needs a model describing the data concepts, and which are filterable/searchable meaningfully.

This idea is also related to which aspects are practical to choose a dataset by, or require access to the internals of a big dataset, and therefore the internal search query needs to be exposed (like the RDA recommendations https://rd-alliance.org/group/data-citation-wg/outcomes/data-citation-recommendation.html )

HTH and doesn't confuse. I will expand further if required.

Chris

-----Original Message-----
From: Andrea Perego [mailto:andrea.perego@jrc.ec.europa.eu] 
Sent: Tuesday, September 27, 2016 8:38 AM
To: Dan Brickley
Cc: public-sdw-comments@w3.org
Subject: Dataset discovery project @ Google (was: Re: Joint session on vocabulary with Web of Things)

(opening a separate thread on this topic)

Thanks for sharing this, Dan.

I have a question about the project's scope:

The objective is mainly focussed on domain-independent dataset discoverability, or it takes into account also domain-specific requirements? E.g., for geo data, the coordinate reference system(s) used in a dataset can be a filtering criterion, but this is not always the case for other data.

Cheers,

Andrea


On 26/09/2016 18:09, Dan Brickley wrote:
> On 26 September 2016 at 17:03, Frans Knibbe <frans.knibbe@geodan.nl> wrote:
> [snip]
>>
> Something to share w.r.t. dataset discovery using schema.org: at 
> Google we are (in the relatively early stages of) exploring what we 
> can do to improve the discovery of datasets. First doc on this was 
> published during TPAC, 
> https://developers.google.com/search/docs/data-types/datasets --- 
> comments welcome here or offlist.
>
> cheers,
>
> Dan

Received on Tuesday, 27 September 2016 18:43:37 UTC