belgif / fedvoc

Federal Vocabularies
6 stars 0 forks source link

Standardize data types / concept schemes for NACE code #30

Open pvdbosch opened 1 year ago

pvdbosch commented 1 year ago

There are different variants. Each additional digit in the code is a more detailed subcategory.

JDMKSZ commented 10 months ago

NACE-BEL 2008: https://statbel.fgov.be/nl/over-statbel/methodologie/classificaties/nace-bel-2008

pvdbosch commented 7 months ago

From what I can find, there are three main contexts in which NACE-codes are used in Belgium:

Examples:

The different variants could be represented like this:

NaceCode:
  description: NACE activity code (international). The first two digits are division of activity. Each further digit specifies an additional level of detail within the category of the previous level. The meaning of the first four digits are standard and defined by EU. Further digits may vary from country to country.
  type: string
  pattern: '^\d{2}\d*$'

NaceBel2008Code:
  description: Belgian NACE activity code of 2 up to 7 digits long. The meaning of the first four digits are standard across all countries. The last 3 digits are defined in the Belgian NACE 2008 list.
  type: string
  pattern: '^\d{2}\d{0,5}$'
  # this is a more constrained version of the international NaceCode schema

NaceBel2008Length5Code:
  description: Belgian NACE activity code of 5 digits long, as used for NSSO and for VAT purposes. The meaning of the first four digits are standard across all countries. The last digit is defined in the Belgian NACE 2008 list.
  type: string
  pattern: '^\d{5}$'
  # this is a more constrained version of the variable-length NaceBel2008Code schema

We could also use allOf to indicate a specialization (NaceCode > NaceBelCode > NaceBel5Code), TBD what the impact would be. More specialized types are less flexible to change, but more clear in which detail level is used and more useful for input validation.

Above types could be included as "DataType".

We also could include three concept schemes:

As for properties, we currently have "economicActivity" http://vocab.belgif.be/ns/other#economicActivity and "orgActivity"http://www.w3.org/ns/regorg#orgActivity; both in draft. They seem to be same thing. I'd prefer to have no "org" prefix to be consistent with other vocabularies, so rather "economicActivity" (or "activity"). Distinguishing between VAT/CBE/NSSO use cases would be going a bit too far IMHO, we'd be importing current administrative complexity in fedvoc.

pvdbosch commented 4 months ago

I tweaked above openapi propositions a bit:

These types can be added to a new openapi schema belgif-openapi-organization.

MarcBruyland commented 4 months ago

Closed as decided in the fedvoc meeting of 2024-06-28.

barthanssens commented 4 months ago

FWIW there has been a NACE2008 skos thesaurus (https://vocab.belgif.be/auth/nace2008?lang=en) available for quite a few years now

pvdbosch commented 4 months ago

@barthanssens , would it be best then to use http://vocab.belgif.be/auth/nace2008 as the URI of the ConceptScheme?

A link could also be added in the comments so users can find the URIs of the individual Nace2008 concepts.

So in NaceBel2008 ConceptScheme of fedvoc:

barthanssens commented 4 months ago

The URI for the skos:ConceptScheme is indeed http://vocab.belgif.be/auth/nace2008
(not http://vocab.belgif.be/ns/other#nacebel2008#id)

pvdbosch commented 4 months ago

OK, would be best to use the existing URI then. @MarcBruyland , could the changes in my previous command be applied to fedvoc?