search

linked-statistics / COOS

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

Statistical or not Statistical #59

Open FranckCo opened 3 years ago

FranckCo commented 3 years ago

Some of COOS classes have "Statistical" in their name (StatisticalProgramCycle), other don't (Task). In some cases it is intentional (distinction between Activity and StatisticalActivity), in some cases maybe less. Can we establish a clear rule that could improve coherence in the naming?

InKyungChoi commented 2 years ago

@FlavioRizzolo @JALinnerud Dan Below is a list of coos classes that has "Statistical" in their name.

Among these, I am least sure about below four:

Additionally, I wonder if we need "Statistical" for "InternationalAgency". Why it is "NationalStatisticalInstitute" but not "InternationalStatisticalAgency" if it is defined as "A body with an international membership, scope, or presence whose primary role is the production of official statistics"? image

JALinnerud commented 2 years ago

StatisticalActivityArea also occurs in the COOS document that was sent out for review.

JALinnerud commented 2 years ago

7 Statistical ios in GSIM v1.2

Business Group 5 ios Statistical Need Statistical Program Statistical Program Cycle Statistical Program Design Statistical Support Program

Concept Group Statistical Classification

Exchange Group Statistical Register

Base Group and Structure Group have zero ios beginning with Statistical..

Revision of GSIM: so far two changes in names (Transformable Input/Output to Core Input/Output)

JALinnerud commented 2 years ago

Notes from meeting 22. February 2022. Participants: Flavio, InKyung, Dan, Jenny coos:StatistcalEntity should be removed everywhere. coos:StatisticalProduct should be replaced by coos:Product. coos:StatisticalDataset should be replaced by coos;Dataset We did not reach a consensus on the others. The GSIM revision is replacing information object with class. For StatisticalOrganization we should also check #65. for GAMSO we should check #10, #63

JALinnerud commented 2 years ago

Notes from meeting 11th of March 2022. Participants: Dan, Flavio, InKyung, Jenny, Nala (a cat) StatisticalActivityArea should be removed. It only exists in the document that was sent out for review. COOS is Core Ontlogy for Official Statistics so we should restrict our scope to official statistics, coos is the namespace and provides the context, but it should be possible to reuse classes in other ontologies/contexts. The provenance of the product and the methodology used in the process are important distinguishing criteria. A possible rule: The adjective satistical in statistical A' should only be used if we can also foresee a use for A without statistical in front eg statistical organisations use both classifications and statistical classifications where categories in the latter must be "mutually exclusive and jointly exhaustive of all objects/units in the population of interest", but not in the former. We could use attributes, definitions, inference (relationships between classes) in some cases rather than the adjective statistical. The discussion will continue in another meeting 29th March CET 16-17 (with or without Nala).

FranckCo commented 2 years ago

StatisticalActivityArea is corrected to ActivityArea in commit https://github.com/linked-statistics/COOS/commit/2ea83bee7ba8dbef436dd93d9d0daadb996d0f90.

JALinnerud commented 2 years ago

Notes from meeting 29th of March 2022. Participants: Dan, InKyung, Jenny. Flavio was unable to attend today. In the absence of Flavio, we found no strong reasons to keep the adjective statistical anywhere, with the possible exception of statistical organisation. The provenance of the product and the methodology used in the process can be distinguishing criteria. Zip codes meet the criteria for a statistical classification (mutually exclusve and exhaustive), but are not designed for statistical purposes.

FlavioRizzolo commented 2 years ago

The other one I think we should keep is StatisticalActivity. We are integrating models that have activities related to statistical production/support, e.g. GSBPM, and others like capability development and strategy & leadership, e.g. GAMSO. They are very different and I think that providing somehow a way of distinguishing between them is necessary. Whether that is achieved by a StatisticalActivity class of by some other means it's a matter to be discussed. I just don't want to lose that information, and it's not something easy to came by from the context because neither model provides a clear cut classification.

JALinnerud commented 2 years ago

Meeting 16th May. (planned) Participants: Dan, Flavio, InKyung, Jenny. Proposal: In COOS, keep the adjective statistical only for statistical organisation and statistical activity.

InKyungChoi commented 2 years ago

I agree with the proposal. Following up questions:

  1. Is it correct we also keep "statistical" in the Statistical Production Activity as it is a sub-type of Statistical Activity?
  2. what do we propose to GSIM revision given that there are terms in COOS and GSIM referring to the same but with different names (e.g., Statistical Program in GSIM - Program in COOS), or is it okay we keep them different?
JALinnerud commented 2 years ago

Meeting 16th May. Participants: Flavio, InKyung, Jenny. Dan did not attend this time. The participants agreed that we only need to keep the adjective statistical for statistical organisation. COOS is the aim. GAMSO and GSBPM in their next revisions should move closer to COOS e.g. cleaning up overlap between overarching processes in GSBPM and corporate support in GAMSO. In the meantime, it is OK that they have different terms. The titles of the models all make it quite clear that the context is and always will be statistical.

FranckCo commented 2 years ago

I would keep "statistical" on both Organization and Activity.