Wikidata:Property proposal/Rotten Tomatoes score

Rotten Tomatoes score edit

Originally proposed at Wikidata:Property proposal/Creative work

   Not done
Descriptionpercentage of professional critic reviews that are positive for a given film or television show
Data typeQuantity
Domainfilm (Q11424) and television program (Q15416)
Allowed values0 to 100
Example 1Titanic (Q44578) → 89
Example 2Avengers: Endgame (Q23781155) → 94
Example 3Nightcrawler (Q15154975) → 95
Example 412 Angry Men (Q2345) → 100
Sourcehttps://www.rottentomatoes.com/
Planned useCreating templates for English Wikipedia to retrieve this data for use in film articles.
See alsoRotten Tomatoes ID (P1258)

Motivation edit

For use on Wikipedia. See https://en.wikipedia.org/wiki/Wikipedia_talk:WikiProject_Film#Rotten_Tomatoes_info_editing_bot. Planning on also creating properties for the review count and the average rating. Similar properties for Metacritic and IMDB might also be useful. Notsniwiast (talk) 15:54, 30 August 2021 (UTC)[reply]

Discussion edit

Here's an example for movie: Titanic (Q44578)
review score
  89%
review score by Rotten Tomatoes
number of reviews/ratings 193 critic review
1 reference


add value
Regards Kirilloparma (talk) 01:57, 1 September 2021 (UTC)[reply]
I agree this is a nice solution, but I can't find a way to include the average rating nicely. The two options I can think of using this solution is to write something like "89%, 7.8/10", or to include another separate value for review score (P444). Do you have an idea for the average rating? Notsniwiast (talk) 04:10, 1 September 2021 (UTC)[reply]
I bet the obvious thing would be to add them as separate values and qualify one or both. determination method (P459) is probably a good qualifier to use, though I don't know what the values of that, in turn, should be. One is a percentage arrived at by pigeonholing all surveyed reviews as positive or negative, and the other is a mean based AFAIK on only reviews that give a score (be it a letter grade, stars, or what have you). Nardog (talk) 09:51, 3 September 2021 (UTC)[reply]
My current idea is just to use separate values and use regex to determine if it is the percentage score or the average. I couldn't find a good value for determination method (P459) that fully describes the RT ratings. Also now that you've pointed it out, you're right that only reviews that give an original score are used for the average, so we shouldn't use number of reviews/ratings (P7887) to qualify the average rating since it's not the same. Winston (talk) 13:33, 3 September 2021 (UTC)[reply]
Actually after some thinking about the technical implementation, it would be much easier to put both scores as a single entry, like "80%, 7.3/10". This also results in less clutter for Wikidata. Winston (talk) 18:54, 3 September 2021 (UTC)[reply]
I've decided to go with Nardog's idea. See Wikidata:Requests for permissions/Bot/RottenBot. Let this property proposal be withdrawn. Winston (talk) 14:42, 4 September 2021 (UTC)[reply]
@Notsniwiast: I still think number of reviews/ratings (P7887) should be used even for an average rating because that's the number RT gives. Nardog (talk) 03:38, 5 September 2021 (UTC)[reply]
But it's given only for the percentage score. The RT FAQ even explicitly says that only reviews with an individual score are used to calculated the average rating.
I think the display of the percent, count, and average on RT also hints at this. The Tomatometer score is always displayed with the review count close beneath. The average score is displayed slightly further away (when you click for more details).
Besides that, adding the review count (say N) to the average rating statement is simply incorrect. People will think that the average was calculated from N individual scores, which is not true (most of the time). Winston (talk) 17:03, 5 September 2021 (UTC)[reply]
@Nardog, Notsniwiast: I was not aware that RT has two scores , I agree this is an additional challenge.
However I would strongly advise against putting both score in a single entry. This makes reusing the data much more difficult − for example, what if I want to query for movies (say, sci-fi movies from the 90s) and order them by average rating?
I would just add a separate value and qualify it − if nothing better comes to mind, then has characteristic (P1552) will do the job − so smth like:
review score
  89%
review score by Rotten Tomatoes
number of reviews/ratings 193 critic review
1 reference
  8.00/10
review score by Rotten Tomatoes
has characteristic <Rotten Tomato average rating>
1 reference


add value
(I will cross-post to the bot request page)
Jean-Fred (talk) 10:03, 9 September 2021 (UTC)[reply]
Oh − I now read through the bot edits and this is exactly what you folks are doing. Well then − perfect! :-) Jean-Fred (talk) 10:05, 9 September 2021 (UTC)[reply]