search

SEMICeu / SDG-sandbox

The SDG Sandbox creates a space for the review of data models produced by WP4 - Data semantics, formats and quality - in the context of the preparatory work for the Single Digital Gateway Regulation.
14 stars 9 forks source link

Level of redundancy and denormalization in the data models #153

Closed psiotwo closed 1 year ago

psiotwo commented 3 years ago

Birth Certificate data model considers Birth as an event, although it does not have specific attributes. On the other hand, the event of issuing the Birth Evidence is not reified.

Depending on the required use-case of the model, I was wondering whether the Birth is of any value in the model (rewiring parent/child to the certificate seems to do the same job).

Were You considering a more normalized data model for evidences? In particular, having a generic --> Evidence -> Issuance -> Event[attributes date, place].

A partial reason for this might have been an aprior alignment of the data model with ISA Core vocabularies (that prescribe e.g. Person.birthDate as an attribute). If this is the reason, maybe direct reference to which elements are interpreted in which external vocabularies (e.g. ISA Core vocabularies), would be beneficial.

Similar questions relate to the other evidence types;

image

cbahim commented 3 years ago

Thanks for the proposition, @psiotwo. However, several member states reported not being able to provide the sex of the parents. How would you tackle this aspect in your re-modelling proposition? The same goes for dateOfBirth.

psiotwo commented 3 years ago

To answer Your question, I would need to know an answer to https://github.com/SEMICeu/SDG-sandbox/issues/157 first :-) .

The reason is that it depends on the goal of the data model which I am not sure about (maybe I missed it somewhere in the documentation). If the data models are "offerings" of possible attributes, then why not keep sex/dateOfBirth with zero min cardinality. An alternative for sex would be to move it to the Child class. For "dateOfBirth" I cannot see any problem - birth events typically won't be recorded for them.

EmielPwC commented 1 year ago

Thank you for your interest and contribution. Please note that this GitHub space is currently not updated (will be soon deprecated), and similar inputs and requests are now handled by the OOTS Helpdesk.

For your information, the current approach for SDG OOTS aims at the reuse of existing data models (where possible) and systems as a possible vehicle for OOTS evidence exchange.

For more information and to stay up-to-date with OOTS developments please consult the recently launched Once Only Hub or reach out to the OOTS Helpdesk.