I think Universe is 'employee income, one component'.
I think Represented Variable is the Measure 'average annual household income'
This would mean only one Instance Variable is needed to define a Data Point
I am unsure if Population of the first cell is supposed to be
a) 2016, Italy
Or
b) 2016, Italy, employee income, one component.
I prefer the latter as the former seems excessive if the Universe is already defined through a represented variable.
Consider making instance variable the specialisation on Universe instead. Represented Variable can then be the measure (or identifier in a unit dataset) and can be reused for different universes and the Universe can have different measures.
On Population, consider linking the attribute Geography to a Classification Item so the Statistical Classification can be the Geography Hierarchy for different NUTS levels.
Reference Period should be a period rather than a date and can similarly be linked to a Category Item.
Even consider normalising Population into Geographic Coverage and Time Coverage as two separate entities.
Note that in Dimensional Data, the Geography and Reference Period represent the overall Coverage (e.g. EU or 2010-2019) where as the Geography and Reference Period of a Data Item would represent the particular Geography or Reference Period of the Datum (e.g. Italy or 2016)
Ygor1970
commented
1 year ago
With the addition of Universe in v1.2, the definition of Population has changed to I’ve a specification of a Universe. In the example of a dimensional dataset in https://statswiki.unece.org/download/attachments/260408186/GSIM%20e-training%20presentation%20-%20Structure%20Group_Scanu_Karling.pptx?version=1&modificationDate=1572518476054&api=v2
I think Universe is 'employee income, one component'. I think Represented Variable is the Measure 'average annual household income'
This would mean only one Instance Variable is needed to define a Data Point
I am unsure if Population of the first cell is supposed to be a) 2016, Italy Or b) 2016, Italy, employee income, one component.
I prefer the latter as the former seems excessive if the Universe is already defined through a represented variable.
Consider making instance variable the specialisation on Universe instead. Represented Variable can then be the measure (or identifier in a unit dataset) and can be reused for different universes and the Universe can have different measures.
On Population, consider linking the attribute Geography to a Classification Item so the Statistical Classification can be the Geography Hierarchy for different NUTS levels. Reference Period should be a period rather than a date and can similarly be linked to a Category Item.
Even consider normalising Population into Geographic Coverage and Time Coverage as two separate entities.
Note that in Dimensional Data, the Geography and Reference Period represent the overall Coverage (e.g. EU or 2010-2019) where as the Geography and Reference Period of a Data Item would represent the particular Geography or Reference Period of the Datum (e.g. Italy or 2016)