Re: [dataset description] Questions to HCLS Community Profile Data Model

Hi Heiner,


On Tue, Sep 12, 2017 at 8:58 PM, Heiner Oberkampf <
heiner.oberkampf@gmail.com> wrote:

> Dear all,
>
>
>
> I've read the Dataset Descriptions: HCLS Community Profile (HCLS-CP) [1]
> and find it very useful. For my application I have the following questions:
>
>
> 1. HCLS-CP recommends 61 properties from 18 vocabularies: Is there a
> merged version available with all those properties in one Turtle file or do
> I need to assemble them on my own?
>
>
> afaik, there is not a turtle file with all the properties.


> 2. void:Dataset is used at the 'Distribution Level'. Would it make sense
> to add two triples to align classes from the different vocabularies:
> (void:Dataset rdfs:subClassOf dcat:Distribution . dcat:Distribution
> rdfs:subClassOf dctypes:Dataset .)?
>
>
> yes, it's too bad that the void file [a] doesn't make any relation between
void:Dataset and the dcat vocabulary. similarly, that dcat [b] isn't linked
to anything.

[a] http://vocab.deri.ie/void.ttl
[b] https://www.w3.org/ns/dcat.ttl


> 3. The CHEMBL example from [1] expresses the distribution URI as "
> http://rdf.ebi.ac.uk/chembl/chembl17rdf"
> <https://clicktime.symantec.com/a/1/1iOHivHIT3TN8B2tOeN7aqB2BMgw4ArjaUQev9Pse1o=?d=iMeKT0dsIJvVr_hhyObCYL__-MJ9GdOoFfepopoMZKumWMxRcCe1II5m5dJ6D2e0TxhlXf0ABLkMxMQqaHToX9URVX3bl3mDAXS9aG4AF6B1F1aRfFi2eV6vbjJKXN0JIV108-FPuCgad9XsNUqibVvkhTnO9LDoLgneu0uHsiiJEA6Qz6dOvV5044seB9gPMesgW5-T9INCWIJGBVbGRkM9mPnDzs-Ns7F-0NTyaqxookCN6il5gGaYNkCyGKPqnmET1JgJL0mukkM5kxe3Q9chgcfefgCFe0gX2brxadqteUng5ne0UiQGXM8KARc5n0R89ehLSOXSzmbkqZOnD_up7DUf7q7TxDBV_x_bxJeebrJpMHKtKTay46GNg-dRlS3u8l5PE6o%3D&u=http%3A%2F%2Frdf.ebi.ac.uk%2Fchembl%2Fchembl17rdf>
> . What are your thoughts regarding a URI pattern such as "
> http://pid.example.com/dataset/{datasetid}/{versionid}/{distributionid}
> <https://clicktime.symantec.com/a/1/3JjJDSkM02gl6B5WL0bKxQmSH_1i_Llp5vlrpzaL9uo=?d=iMeKT0dsIJvVr_hhyObCYL__-MJ9GdOoFfepopoMZKumWMxRcCe1II5m5dJ6D2e0TxhlXf0ABLkMxMQqaHToX9URVX3bl3mDAXS9aG4AF6B1F1aRfFi2eV6vbjJKXN0JIV108-FPuCgad9XsNUqibVvkhTnO9LDoLgneu0uHsiiJEA6Qz6dOvV5044seB9gPMesgW5-T9INCWIJGBVbGRkM9mPnDzs-Ns7F-0NTyaqxookCN6il5gGaYNkCyGKPqnmET1JgJL0mukkM5kxe3Q9chgcfefgCFe0gX2brxadqteUng5ne0UiQGXM8KARc5n0R89ehLSOXSzmbkqZOnD_up7DUf7q7TxDBV_x_bxJeebrJpMHKtKTay46GNg-dRlS3u8l5PE6o%3D&u=http%3A%2F%2Fpid.example.com%2Fdataset%2F%7Bdatasetid%7D%2F%7Bversionid%7D%2F%7Bdistributionid%7D>"
> , to express the three levels (summary/version/distribution) also within
> the URI?
>
> seems like a reasonable pattern to me!

m.


> Any suggestions are very appreciated.
>
>
> Kind regards,
>
> Heiner
>
>
>
> [1] https://www.w3.org/TR/hcls-dataset/
>



-- 
Michel Dumontier
Distinguished Professor of Data Science
Maastricht University
http://dumontierlab.com