4

I would like to know what fraction of the recent close-vote review tasks were initiated by a flag cast by a user under 3k reputation (as opposed to tasks assigned to the queue by a close vote by a 3k+ user). Is there enough data in the Data Explorer to get this information? If so, how can this be done?

There isn't any particularly obvious direct path for this, so I thought of getting the first user that cast the close vote and then matching this against the users that reviewed, since those are public (example), but it seems that the ReviewTaskResults table does not include any information on the users that performed the reviews, for whatever reason.

1

There is a list of pending flags, but sadly that obviously doesn't have historical data.

Besides that table, there is no information in SEDE regarding if a post was flagged and for what reason. Only public information like who were responsible for closing it is shown. (As a note: if I vote to close, the vote can expire when no one adds a vote to it. After that it can be closed, but my name won't be on it)

  • I don't care about who flagged a post - I want to know if it was flagged or not. The only name information that would be useful is the name of the first reviewer (which matches the first closevoter if flag, doesn't match if direct vote) but this is not the name of the flagger. – E.P. Jul 18 '16 at 12:46
  • I also don't care about pending flags - I want historic data, say, older than last week. – E.P. Jul 18 '16 at 12:46
  • I should have written my post a little better @E.P. Corrected. – Patrick Hofman Jul 18 '16 at 12:48
  • Again, note that I can make do with public information, since the names of the reviewers are public (see the linked example), but that public information is not present in the ReviewTaskResults table for some reason. (The case in your parenthesis also strikes me as an infrequent edge case, and one that would also be naturally well handled w.r.t. the information I want to get.) – E.P. Jul 18 '16 at 12:53
  • Flags aren't public information... Close votes are. You can't know if the item was added through a flag, some automated system, etc. – Patrick Hofman Jul 18 '16 at 12:54
  • If there is some automated system, I'm happy to lump that in with flags (but I'd be curious to know what it is and where it is documented). Please take this question seriously and accept that I have thought about this already; I only require public information. The names of close-voters are public and the names of reviewers are public. If the name of the first close-voter is not in the list of reviewers, then the review task was created by the first close-voter's cote. If the first close-voter is in the list of reviewers, it was some other mechanism. – E.P. Jul 18 '16 at 12:59
  • Review task by review task, the information I want is public. Now I just want the stats. – E.P. Jul 18 '16 at 12:59
  • 1
    Keep in mind that on high traffic sites posts might get closed without ever seeing a review task. But that said: It doesn't sound unreasonable to have historical raised close flags and votes against posts – rene Jul 18 '16 at 13:15
0

You can't query that historical data in SEDE because the flags raised against a post are not part of the public schema.

You can get some information if you query pendingflags:

;with basedata as
(
select sum(case when flagtypeid = 13 then 1 else 0 end) as closeflags
     , sum(case when flagtypeid = 14 then 1 else 0 end) as closevotes
     , postid
from pendingflags
where flagtypeid in (13,14)
group by postid
)

select (
  select count(*) 
  from basedata 
  where closeflags = 0
  ) / 
  cast( (select count(*) 
         from basedata  
         where closeflags = 1 and closevotes > 0) 
  as decimal)

Flagtypeid 13 indicates a close flag against a postid where flagtypeid 14 indicates a real close vote. Unfortunately the table pendingflags only hold data while a post is in the review queue. Once completed you're only left with the close vote data in the posthistory table for rows with posthistorytypeid = 10. That doesn't include if the closing involved close flags.

This is the best I can create on the current schema. Consider opening (or change your question in) a feature request to add a pendingflagshistory if you need the history to be available.

Or you might ask Shog9 to re-run his query for you on the internal SEDE instance.

  • No, I do care about historical data - I only care about completed reviews. As I have said in the question and in the comments to Patrick's answer, the information I want is public. I want to know if it is available through the data explorer. – E.P. Jul 18 '16 at 13:05
  • 1
    @E.P. I was afraid you would ;) Maybe you need to elaborate a bit in your question what the use case is for that data. That will help if they consider that to be added in the next 6 to 8 .... – rene Jul 18 '16 at 13:11
  • The use case for the data is on the question title and first sentence. (In case it's not obvious, it has a direct bearing on knowing how much the 15-3k rep user segment is active in close-related community moderation.) – E.P. Jul 18 '16 at 13:13

You must log in to answer this question.

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