Wikidata:Property proposal/number of people depicted

number of people depicted

Originally proposed at Wikidata:Property proposal/Creative work

   Withdrawn

Description	people depicted in an image, such as a painting. Can be used to check completeness values in "depicts" (P180)
Data type	Item
Domain	items for works such as paintings, photographs
Allowed values	items for numeric values (e.g. "1", "2", "3", ">8"); not items for persons
Example 1	Mona Lisa (Q12418) → 1 (Q199)
Example 2	The Balcony (Q775407) → 4 (Q202)
Example 3	The Last Supper (Q128910) → 13 (Q37141)
Example 4	The Bull (Q2917717) → 1 (Q199)
Example 5	Portrait of a Woman with a Squirrel (Q17335769) → 1 (Q199)
Example 6	The Last Supper (Q128910) → ">8" based on heuristic (P887) TensorFlow (Q21447895) (fictitious sample, per request below)
Robot and gadget jobs	Could be used for complex constraint on P180, Sample query: [1] Could be added by bot for manual review
See also	depicts (P180), relative position within image (P2677)

Motivation

See applications above
--- Jura 04:09, 8 July 2018 (UTC)

Discussion

•  Support. @Jura1: Is not it better to use datatype "quantity"? Thank you David (talk) 15:11, 8 July 2018 (UTC)
  • @ديفيد عادل وهبة خليل 2: I hesitated about that. A bot could add an item that just says ">8". It might be harder to express/query it with quantity-datatype. What do you think?
    --- Jura 15:22, 8 July 2018 (UTC)
    • @Jura1: Is it possible to mention an example of the reason for the hesitation? Thank you David (talk) 15:37, 8 July 2018 (UTC)
      • @ديفيد عادل وهبة خليل 2: I added a sample.
        --- Jura 15:50, 8 July 2018 (UTC)
        • @Jura1: Why do we need marks? We can make bots only add quantities without marks David (talk) 15:56, 8 July 2018 (UTC)
      • @ديفيد عادل وهبة خليل 2: You meant the quotes? It's just to indicate that it's the label of the item. Currently, there is no item for more-than-eight (>8) and I didn't want to create one just for discussion purposes.
        --- Jura 16:00, 8 July 2018 (UTC)
•  Oppose use "depicts=Q5" (or, as in the case of The Balcony (Q775407), man, boy, named individuals, etc.) with qualifier "Quantity=[n]" if more than one. And items for numeric values are most certainly not how to represent quantities. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:37, 8 July 2018 (UTC)
•  Support @Cwf97: 18:31, 9 July 2018 (UTC)
•  Oppose, per Andy Mabbett. --Yair rand (talk) 01:55, 12 July 2018 (UTC)
  • @Yair rand: I can't make any sense of it, can you suggest a query that checks that the enumeration in "depicts" is complete? As I know Pigsonthewing doesn't write any SPARQL, I can't ask him.
    --- Jura 07:49, 12 July 2018 (UTC)
  •  Comment I asked for input on request a query as well, but given the lack of feedback, it seems the suggested alternative isn't possible. Accordingly, I'm marking this as ready.
    --- Jura 04:48, 15 July 2018 (UTC)
•  Oppose: Andy's solution is much more generic. And the item datatype is puzzling - why would you prefer that over quantity? − Pintoch (talk) 08:11, 17 July 2018 (UTC)
  • @Pintoch: please see above. If there is a preference for quantity datatype, we could change that. What is your suggestion for querying completeness? I'd would be good to have a SPARQL sample. Unfortunatly, we haven't had an suggestions yet.
    --- Jura 11:01, 24 July 2018 (UTC)
    • @Jura1: your request for a query states that there is a suggestion to use "depicts" recursively to check its own completeness. Is this doable? I don't understand where you see any recursivity here. Can you describe in plain English what you want to see as the output of your query? − Pintoch (talk) 08:29, 25 July 2018 (UTC)
      • I added a sample query above that would work with this property. The idea is to do the same only with P180. When preparing the sample query, I noticed that I can't even to determine reliably the number of people from P180, but maybe I'm just using it wrong. Can you do it?
        --- Jura 09:17, 25 July 2018 (UTC)
        • I would just replace ?sample wdt:P369 / wdt:P1181 ?qt by ?sample p:P180 [ ps:P180 wd:Q5 ; pq:P1114 ?qt ]. You would need to change your wdt:P31* in the subquery to wdt:P31 so that this claim is ignored in the subquery. Maybe we actually need a property "depicts objects of class" to make the distinction between "depicts human" and "depicts Douglas Adams", that would solve your "recursivity" issue. − Pintoch (talk) 11:04, 25 July 2018 (UTC)
          • @Pintoch: That's not the way "depicts" is being used nor is it what you supported above ("man, boy, named individuals, etc."). The samples I added above all come from properties, so I think they are representative for the way P180 is meant to be used. Why the reluctance to create a new property? I don't see a downside. People not interested in it could just skip it.
            --- Jura 11:11, 25 July 2018 (UTC)
          • @Pintoch: BTW, could you create a query that matches your proposed query ("You would need to change ..")? I'm not sure if it's even possible without timing out.
            --- Jura 11:32, 25 July 2018 (UTC)
            • @Jura1: I'm not sure why you can't just do the changes I described, but there you go: tinyurl.com/ycj4pl2x − Pintoch (talk) 08:51, 29 July 2018 (UTC)
•  Comment I do agree that we should just use quantity (P1114) as a qualifier and that the proposed datatype is strange whatever. Thierry Caro (talk) 17:56, 17 July 2018 (UTC)
  • @Thierry Caro: What is your suggestion for querying completeness? I'd would be good to have a SPARQL sample.
    --- Jura 11:01, 24 July 2018 (UTC)
•  Oppose, per Andy Mabbett. Relf PP (talk) 06:24, 18 July 2018 (UTC)