Re: Brainstorming data quality needs within the Dataset Usage Vocabulary.

Antoine,

Thank you for your responses and the clarification, I've added my comments
to yours.  With the exception of the question about how to depict the
qualifications of someone's opinion I think the Dataset Usage Vocabulary
has a viable path forward.

Cheers,

Eric S

On Sat, Apr 18, 2015 at 3:19 AM, Antoine Isaac <aisaac@few.vu.nl> wrote:

> Hi Eric,
>
> (I'm evacuating the DCAT issue - to me it's not really worth discussing
> here, it's more a meta-requirement and anyway we already said that we
> should base our work on DCAT...)
>
>
>  I was thinking that the data quality vocabulary would support objective
>> metrics, subjective metrics, and qualified opinions on the dataset as well.
>>
>>
>
> I agree that the Q&G voc should allow for opinions/rates, this is already
> in our requirements. What I disapproved with was to include in scope right
> now the idea of 'rating the raters'. This would lead us way too far, while
> we're struggling to capture basic requirements.
>
> +1  okay I understand your perspective.

>
> On your specific points:
>
>  Proposed example objective metrics on dataset:
>> * Consumer DCAT:Dataset usage should be tracked by metrics such as a
>> counter.
>>
>
>
> This looks like something for the data usage vocabulary.
>
> +1 okay sounds good

>
>> Proposed example subjective metrics on dataset:
>> * Metrics should be used to rate consumer acceptability for a
>> DCAT:Dataset.
>>
>>
>
> This could be for the quality vocabulary. Well at least for the fact that
> consumer could express opinions the dataset. I have no clue yet how these
> opinions would result in metrics.
>
> +1 I could see for instance using similar approaches as they evolve in the
data quality vocabulary.  I'd like to aim for being synergistic or
complimentary approaches when possible.

>
>
>> Proposed example qualified opinions:
>> * Metrics should be used to rate qualifications of consumer providing
>> opinions about a DCAT:Dataset
>>
>>
>
> The wording is unclear to me, but this reads like 'rating the raters' so
> would be out-of-scope.
>

The question I have in my mind here is "what is qualified opinion?" is it
conveying credentials in some fashion?  Perhaps rating someone is too
simplistic of a model for both vocabularies.

>
> cheers,
>
> Antoine
>
>
>
>
>
>>
>>
>> On Thu, Apr 16, 2015 at 1:56 PM, Antoine Isaac <aisaac@few.vu.nl <mailto:
>> aisaac@few.vu.nl>> wrote:
>>
>>     Hi Eric,
>>
>>     This is quite an interesting discussion...
>>
>>     My two cents would be that the following would be in scope for the
>> Quality vocabulary
>>     "Consumers should be able to provide feedback on overall DCAT:Dataset
>> quality"
>>
>>     The rest would be out-of-scope. Maybe from the perspective of the
>> data usage vocabulary it make sense to further qualify the 'raters', in the
>> case they would be also data users.
>>
>>      >From the perspective of quality voc that's just a no-go. We're
>> fighting to get concrete requirements for a framework for representing
>> quality of datasets, we shouldn't embark on getting a framework to
>> represent the quality of people.
>>
>>     Cheers,
>>
>>     Antoine
>>
>>
>>     On 4/16/15 10:47 PM, Eric Stephan wrote:
>>
>>         Question:  Does the following help clarify/confuse the quality
>> needs from the Dataset Usage Vocabulary perspective?
>>
>>
>>         I'm not sure if anyone from the Data Quality Vocabulary was on
>> the call on the second day of F2F3 when we discussed the Data Usage
>> Vocabulary topic.
>>
>>
>>         I think there were some important points that were made:
>>
>>         1) The "Data Usage Vocabulary" had been changed to the "Dataset
>> Usage Vocabulary".
>>
>>
>>         2) This name change was done with the intention to focus our
>> efforts on providing a vocabulary at the DCAT:Dataset level only.
>>
>>
>>         3) By focusing efforts at the DCAT:Dataset level it allowed us to
>> avoid the seemingly endless are we talking data or dataset discussions.
>> Most importantly it allowed us to talk about DCAT:Dataset has being a
>> logical container for a "set of data".
>>
>>
>>
>>         My hope is that this might simply how we build bridges between
>> the Dataset Usage Vocabulary and the Data Quality Vocabulary.
>>
>>
>>         Below are listed some tangible minimal quality related
>> requirements for DCAT:Dataset Usage:
>>
>>
>>         * Consumer DCAT:Dataset usage should be tracked by metrics such
>> as a counter.
>>
>>         * Consumers should be able to provide feedback on overall
>> DCAT:Dataset quality
>>
>>         * Consumers should be rated for their qualifications when
>> commenting on DCAT:Dataset.
>>
>>         * Metrics should be used to rate consumer acceptability for a
>> DCAT:Dataset.
>>
>>                        -   Boolean metrics should be used to indicate
>> overall approval/disapproval.
>>
>>                        -  Integer scale metrics should be used to
>> indicate levels of acceptability.
>>
>>         * Metrics should be used to rate qualifications of consumer
>> providing opinions about a DCAT:Dataset
>>
>>
>>         I welcome more ideas for requirements, but these requirements do
>> illustrate the data quality needs of the Dataset Usage Vocabulary.
>>
>>
>>         I look forward to your thoughts and ideas,
>>
>>
>>         Eric S
>>
>>
>>
>>
>>
>>
>