Re: health care schema

On 6 December 2013 08:45, Bradley, Lily (HHS/ASPE) <Lily.Bradley@hhs.gov> wrote:
> Thanks all! Very helpful. If I were looking at a dataset about hospitals,
> what's the 'strongest' (I.e., most "linked") way to describe the content
> of that dataset?
>
> Eric, Is this sufficient/useful for consuming Medicare hospital compare
> data:  https://data.medicare.gov/developers ? Other priorities for new
> APIs?
>
> Simon, The US hospital characteristics file is also available via API:
> https://data.cms.gov/d/xssf-xjt2?Category=Medicare&view_name=Provider-of-Se
> rvices-File-OTHER-September-2013 The data dictionary is zipped on cms
> website, but available as text here:
> http://www.nber.org/pos/2011/POSQ.OTHER.LAYOUT.DEC11.txt
>
> Dan, Scott's in town next week. Which property would "hospital" best fit
> into here: http://schema.org/Dataset

We have a specific type for http://schema.org/Hospital ... whereas
http://schema.org/Dataset is (currently) a kind of black box container
for arbitrary data.

There's isn't yet a good bridge between these two views of data.
Either you express specific facts using types/properties, or you
provide (say) CSV data files e.g. zipped, and call the whole bundle a
'Dataset'. But we're looking at ways of bringing things together, so
that rows in a CSV data file that happen to describe the properties of
(say)  a Hospital can be exposed as such. There's a paper at
http://www.w3.org/2013/04/odw/odw13_submission_53.pdf talking a bit
more about this, and a proposed W3C group on tabular data,
http://www.w3.org/2013/05/lcsv-charter.html so I'm interested to
collect practical use cases. I'll take a look at the links above...

cheers,

Dan



> Joanne and Tim, I will continue to explore those references. Thanks so
> much!

Received on Friday, 6 December 2013 08:53:32 UTC