Project:Vocabulary/GDPR

From Wikibase Personal data
< Project:Vocabulary
Revision as of 22:45, 11 February 2020 by Podehaye (talk | contribs) (→‎Issues/Clashes)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Use cases

The Ontology need to be suitable for data entries and the usages of those data. We need to identify use case and test the ontology against those.


Examples

Core snaks

  • instance of (P3) - data controller (Q96)
    • Both GDPR generators will not pop up if this is not present.
  • e-mail address (P17) - "mailto:johnny@example.com"
    • for (Q1062) - data controller (Q96)
      • This qualifier is not currently checked but it should be, as there may be a general support email as well as a data protection officer specific email address and we need to pick the correct one.
  • contact page (P282) - "https://company/privacy/contact"
    • This is necessary where there is no known email address for contacting the data protection officer, only a form. In these cases you probably still need to add e-mail address (P17) but with a dummy email address (and a comment qualifier making clear it is not a real address). The generated email then has to be manually pasted into the contact form for submission. For example see Apple (Q3424)
  • collects (P10) - the points of data known to be collected/stored by this company.
    • named as (P27) - "alternative string name"
      • This qualifier allows existing items to be re-used. The idea is that the 'named as' value is the display name that is used in the GDPR request letter. So if we already have 'location' but want to include 'geolocation info' because that is the term used for this concept from the company's privacy policy, we can include location (Q1078) in the collects (P10) list but add named as (P27) - 'geolocation info' as a qualifier. This is important to allow the eventual unification of different data concepts into one model despite companies describing them differently. This is currently only used in Alex Bowyer's GDPR Generator (User:Alexbfree/Code:GDPRMailAccess.js). It is recommended this be merged to the Le Temps generator. Another use of this field is to convert the pronoun to the correct form for the letter - for example the stored data point might be "information about your journeys". You can add a named as (P27) to change this to "information about my journeys" in the generated GDPR letter.
    • source (Q1117) - privacy policy (Q29) or source (Q1117) - privacy hub (Q3732)
      • This qualifier is especially important where other data points have already been added to the collects (P10) list from different sources. This is used to distinguish things we have learnt from different sources. We should update our GDPR generators to only include the collects (P10) items from the sources we want to use (which may differ per project).
    • policy excerpt (P287) - "Extract from privacy policy which shows why this data point is believed to be stored."
      • Add more than once if a long extract is needed. Take care to remove spaces from end of string or you will get a `Malformed text` error.
  • official website (P15) - URL (optional really)
  • states compliance (P32) - privacy policy (Q29)
  • requires (P122) - any item that is required by the company
    • Currently all items in the requires (P122) list will be included at the end of the GDPR letter as possible things that the data subject will provide to identify themself. We probably need to refine this - to differentiate between things that are needed to even identify a customer to kick off a request (account number, username etc.) vs things that may be asked for by the company after a GDPR request has started (e.g. security checks, copies of proof of address etc).
    • named as (P27) - "alternative string name"

Issues/Clashes

[Template:For] is a template that essentially applies one template in succession to a bunch of parameters, with a separator (ignore the named variables)
this is based on the module [Module:For]
do you understand why the template is limited to five named parameters while Module:For does not explicitly implement such limitation?
in any case what you want to solve your problem is a module TwoVariableFor, which takes a bunch of parameters and feeds them two at a time into a rendering function that takes a pair of arguments, the second one being possibly empty
then you want to have something like {{#invoke:TwoVariableFor|call=MyRenderingTemplateInPairs|firstitem|firstqualifier|seconditem||thirditem|thirdqualifier}} (( for instance the second qualifier was missing)
now the question is how you generate the list of parameters, but you can use an invoke:Wd for this
so: {{#invoke:TwoVariableFor|call=MyRenderingTemplateInPairs|{{#invoke:wd|...what you already know how to do...}}}}
it would be awesome if you managed to duplicate and modify Module:For into Module:TwoVariableFor

FAQ

Q: how do we refer to privacy policy? there are a few different ways to do it.

  • A: The correct one is privacy policy (Q29) though we have previously used Q3405 and some usages of this still need to be migrated. Q3667 and Q173 are erroneous and should be deleted.

Data rights snaks

Multiple choice between different options

References

PREFIX pdio: <https://wiki.personaldata.io/entity/>
PREFIX pdiot: <https://wiki.personaldata.io/prop/direct/>
PREFIX pdiop: <https://wiki.personaldata.io/prop/>
PREFIX pdiops: <https://wiki.personaldata.io/prop/statement/>
PREFIX pdiopq: <https://wiki.personaldata.io/prop/qualifier/>
SELECT ?item ?itemLabel ?link WHERE {
  ?item pdiot:P110 pdio:Q3006. #P110 <-concerns Q3006 <-ontology model	
  OPTIONAL {?item pdiot:P15 ?link.}
  SERVICE wikibase:label {
    bd:serviceParam wikibase:language "en" . 
  }
}
LIMIT 100

sources on (GDPR) ontology modeling (embedded)