search

linked-statistics / COOS

Core ontology for official statistics
Creative Commons Attribution 4.0 International
5 stars 5 forks source link

Feedback from Rosemay McGrath #72

Open FranckCo opened 2 years ago

FranckCo commented 2 years ago

Here are the comments sent by Rosemary:

COOS - RMG feedback.docx

FlavioRizzolo commented 2 years ago

Regarding: We have one question in relation to event data. The “Key-value data are often taken from scraping the web. Even-history (transposed) is used to describe events over some time period” suggests temporal, but not bitemporal?

We haven't consider time explicitly at all in COOS, but I think we need to look at bitemporal at least, i.e. valid/effective vs. transaction/system time. Especially since we are including event data, which suggests some sort of temporal semantics, as pointed out by Rosemary above.

FranckCo commented 2 years ago

All of Rosemary's comments reproduced below:

Stats NZ supports the feedback from Thérèse Lalor and Jenny Linnerud and have not replicated any feedback covered in those documents.

Draft COOS has good coverage of the things to observe and how they relate. We support the work undertaken to develop the core concepts of the Ontology. In particular we can see utilisation of the Ontology as a bridging Ontology to lower level domain ontologies or the development of these.

As you may know Stats NZ has another functional leadership role in the NZ Government – Government Chief Data Steward, and we see that the fundamental concepts can apply more broadly across the NZ Government Data System, for those who produce official statistics and for Data management and Data analysis more broadly.

We particularly like the linking to international statistical standards and international vocabularies that span domains and are extensible to the specifics (potentially country specifics) where required – for example in New Zealand this may be indigenous considerations.

We support the next steps proposed and particularly like the further reference of the EIF as this is a model we have been exploring to support the data system design architecture work.

We have one question in relation to event data. The “Key-value data are often taken from scraping the web. Even-history (transposed) is used to describe events over some time period” suggests temporal, but not bitemporal? And one question which we think we know the answer to – but wondered if you could provide more clarity across the operational versus statistical lenses. Is there any intention to further the operational lenses?

ChLaaboudi commented 2 years ago

Link to EIF https://joinup.ec.europa.eu/collection/semantic-interoperability-community-semic/adms

FranckCo commented 2 years ago

@ChLaaboudi I think this is the link for ADMS (which can be useful if we want to register COOS as a semantic asset). The EIF is described here. The actual document is lost somewhere in limbo as the EC has changed all its links (again)...

FranckCo commented 2 years ago

I had a mail exchange with Rosemary: she confirms she was referring to the European Interoperability Framework.

ChLaaboudi commented 2 years ago

https://joinup.ec.europa.eu/collection/nifo-national-interoperability-framework-observatory/european-interoperability-framework https://joinup.ec.europa.eu/collection/nifo-national-interoperability-framework-observatory/3-interoperability-layers#3.5

ChLaaboudi commented 2 years ago

A new version of EIF, EIRA (5.0.0), was published on 11/5/2022 https://joinup.ec.europa.eu/collection/european-interoperability-reference-architecture-eira/about#eia https://joinup.ec.europa.eu/collection/european-interoperability-reference-architecture-eira/solution/eira/release/v500