RE: DQV - metrics related to the completeness dimension

+1 to this.


On the objective level, a data provider can say that the data includes all 20 air quality parameters measured for 50 stations, and therefore is complete within its own stated context


On the subjective level, the data provider can boast that this represents “high-quality” data, while a user may have the opinion that the 20 parameters are not enough and that the measuring stations are too far apart, and therefore the data is of low quality.


Those are two perspectives that DQV needs to encompass, in my opinion.




From: Laufer [] 
Sent: 30 September 2015 16:01
To: Steven Adler <>
Cc: Nandana Mihindukulasooriya <>; Debattista, Jeremy <>; Data on the Web Best Practices Working Group <>; Makx Dekkers <>
Subject: Re: DQV - metrics related to the completeness dimension


Hi, All,

This issue about "what is quality" was also raised during the F2F SP.

I want to give my opinion (my opinion has my quality point of view):

I  think that what we are providing with the BPs, and the DUV and DQV, are ways for the publishers to aggregate metadata to the published data to help the consumer to discover and decide if a particular data is of her interest and, in this case, to use it.

Of course, the point of view of all descriptions is the point of view of the publisher, but this does not mean that the information is not valid.

In terms of the quality aspects (or characteristics, dimensions, etc.), for each one of them, there must be a clear description of what that aspect is asserting. With this clear description, I do not see why it is a problem to a publisher asserts some quality aspect about the published dataset. The set of assertions will be one more thing to support the decision of the user.

I know this could be not a good example, but when we have the labels in a snack product, with all that information about the composition, etc., it is a decision of each consumer to choose if that snack is ok or not for her. But in the label we have a set of quality aspects (or dimensions, etc.) that can support that decision.



Received on Wednesday, 30 September 2015 14:11:28 UTC