9

This question already has an answer here:

My favorite SE site gets a lot of bad questions, and a large amount of them get closed, it got me wondering if there were other SE sites that get as many bad questions, or if my chosen favorite just attracts more than it's fair share of stinkers. Are there any analytics for keeping track of how many questions each SE site closes and deletes?

marked as duplicate by gnat, PolyGeo, Ward, Nathan Tuggy, Robert Longson Jul 27 '17 at 6:24

This question has been asked before and already has an answer. If those answers do not fully address your question, please ask a new question.

  • 8
    Lots of closed question get deleted, so any data you get will be incomplete because it will be based only on the volume of undeleted closed questions. – Infinite Recursion Oct 25 '14 at 5:32
  • 2
    I'm not really sure why this is receiving so many down-votes. It's not a question we--the users--can answer, so it'll only get an answer if a CM or other employee decides to humor it (I assume the private data explorer would allow for a query such as this), but I still think it's an interesting question, regardless of its answerability. Of course, the answer would also tell you more about the community and VTC culture than the quality of questions they receive, so you'd have to take it with a grain of salt, but that'd still be interesting to learn. – Matthew Haugen Oct 25 '14 at 19:01
  • Can those with a high enough reputation not see deleted questions? Why is there an analytics tag unless you can ask analytical questions? Naturally I expect the answer will come from someone with a high enough reputation or sufficient privileges to view SE analytics. – ShemSeger Oct 26 '14 at 2:13
  • 1
    While I'm sure this is answerable, I can't see it being useful, as the answer will not tell you anything about the relative quality of questions posted on the sites. – Rory Alsop Jan 7 '15 at 15:43
  • Is this idle curiosity, or do you see some kind of action arising, depending on the answer? – EnergyNumbers Jan 7 '15 at 15:49
  • @EnergyNumbers Mostly curiosity, but it would be nice if it could lead to some sort of discussion about attracting better quality questions to those sites that seem to get more off-topic questions than others. – ShemSeger Jan 7 '15 at 16:14
  • 2
    Neither high-rep users nor mods can answer this questions (mods only have access to such stats for the site they moderate). But you can find the answer for the deletion part at Around how many questions get deleted in Stack Exchange and by who?, slightly dated. ... And for non-deleted closed questions you can find the answer yourself, by searching for closed:yes on each site (one can also write a cross-site SEDE query to automate this). – user259867 Jan 7 '15 at 16:53
  • Don't suppose anyone feels like writing a cross-site SEDE query is worth a 150rep bounty? Is there a way you can search for only deleted questions? – ShemSeger Jan 7 '15 at 18:27
  • Make that 200rep, I've got another 50 now that I can reward the accepted answer. – ShemSeger Jan 7 '15 at 18:33
  • You won't have such an ability, cuz the bounty amount doubles with the every next bounty, until it reaches 500. – nicael Jan 8 '15 at 13:13
  • Well then, guess I'll just have to save up. – ShemSeger Jan 8 '15 at 13:25
  • @Gnat, except my question was asked more than 2 years before that one. If anything, it's a dupe of my question, or should have been an answer to my question instead of being asked as it's own question. – ShemSeger Jul 26 '17 at 22:04
  • This seems like a bad question to me. What is a bad question to you may not be to some beginner, or newbie on his way... – dyasta Jul 26 '17 at 23:32
5

With the introduction of the table postswithdeleted this question is now answerable with the following multi-database query:

-- result table, don't rename and keep the site column
create table #results ( site nvarchar(250)
                       , qt int
                       , q int
                       , qct int
                       , dq int
                       , cq int
                       , dcq int);

declare @sql nvarchar(max) = ''   -- holds build up sql string

-- build one biq union sql, for each db
select @sql = @sql 
+ iif( len(@sql) > 1 
     , 'union'
     , 'insert into #results'
) +
-- here goes the per site query, fully qualify the database objects
N'
select ''' + name + '''
     , count(*) qt
     , sum(case when deletiondate is null then 1 else 0 end) q
     , sum(case when closeddate is not null then 1 else 0 end) qct
     , sum(case when deletiondate is not null then 1 else 0 end) dq
     , sum(case when deletiondate is null and closeddate is not null then 1 else 0 end) cq
     , sum(case when deletiondate is not null and closeddate is not null then 1 else 0 end) dcq
from ' + quotename(name) + '.dbo.postswithdeleted
where posttypeid =1
'
from sys.databases
where database_id > 5
-- and (name not like '%.Meta' or name = 'StackExchange.Meta')

--print @sql

-- execute it
exec (@sql)

-- show results
select site
      -- rest of columns
      , qt as [al questions]
      , q as [visible questions]
      , qct as [all closed questons]
      , dq as [deleted questions]
      , cast(qct as decimal) / qt as [c/q ratio]
      , cq as [closed questions]
      , dcq as [delete closed questions]
from #results
order by cast(qct as decimal) / qt desc

drop table #results

As I don't think absolute numbers will be interesting I ordered the result by the closed/questions ratio (overall, so both deleted and visible posts).

When run today this will be the result:

enter image description here

And Programmers (now called Software Engineering) is the winner(?)!

Keep in mind SEDE is updated only once a week, on Sunday.

  • 2
    And the runner up is the site I was alluding to in my question. That answers that. – ShemSeger Jul 26 '17 at 20:51
  • SEDE using outdated sites names is a known issue. All-time stats are skewed by various historical cleanups, eg in 2016 these made close rate over 90% at Anime.SE. At (ex-)Programmers, skewing factor is likely cleanup of inappropriate questions that were mass-migrated from Stack Overflow in 2010-2012. Probably more accurate picture was provided in 2016 stats post (except for Anime.SE which had "organic" close rate broken by some historic cleanup) – gnat Jul 26 '17 at 22:11

Not the answer you're looking for? Browse other questions tagged .