15

This question is an exact duplicate of:

Can someone with access to the right data produce some analysis of the Close Vote queue?

When I work out how, or when I can do it, there will be a substantial (for me) Bounty for something effective.

I want what many want: which is how to deal with the Close Queue. Here are some analysis that may be useful:

  • An analysis by age would be useful.
  • Analysis by tag would be useful.
  • If the "skip count" is available, analysis by that (say three broad groups to begin with - what are the "thirds" of the total queue?).
  • An analysis of users by some grouping of rep, with Moderators another group.
  • An analysis of top users down (to get an idea of how many users need to be encouraged such that an impact can be expected judged by the ratios of work done by people currently).

As well as lots of other things. Having never even seen the queue myself, it is difficult to be definitive or correct. Feel free to edit in suggestions....

Someone in the comments asked whether offering fries would be an incentive. Don't know where to start with that one.

Some more suggestions based on the comments:

  • For each language tag: the ratio of questions in the queue to close-voters with expertise (say, a bronze badge in the tag). SO could recruit more mods from areas in which the ratio is high; and concerned folks could start meta or chat conversations to come up with new norms for questions in those areas. – @Frank

  • Perhaps to extend to get some idea of the potential resource available per language tag (if languages themselves can be identified from tags?). How many, broken down by 3k, 10k, 20k have N+ Accepted answers in language tag.

  • An analysis of 'active' 3k+ users and how many close votes they cast per day on average (in the last ... month? week? year? and in total) (and similar analyses of users at higher reps, e.g. 10k+, 20k+) – @Dukeling

This might underpin, or undermine, the idea that the size of the Close Vote queue is demotivating.

  • Do reviewers slow down once they've achieved a badge milestone?
  • Can the effects of existing robo-voting be quantified?
  • Ratio of review tests passed vs review test failures, for 3k, 10k, 20k.
  • What is adding to the queue the most? In the last 30 days, how many questions are from the last thirty days, and how many from 60, 90, 180, 360, 720 days?
  • How many questions on the queue have (mutually exclusive): Accepted answers; More than one Upvoted answer; One Upvoted answer; More than one answer; One answer; no answers.
  • Same answer as previous, but for events which happened since question became Hold/on the Close Vote queue.
  • Ratio of questions per tag as a whole vs questions per tag on Close Vote queue.
  • On Hold must feed into the Close Votes. Can On Hold movements be analysed for the existing data?

marked as duplicate by animuson Nov 21 '13 at 16:14

This question was marked as an exact duplicate of an existing question.

  • 2
    What kind of analysis do you want? I have pending items by tag (and some variations on that) for an answer I'm working on elsewhere; it's also easy enough for me to break it up by close reason, look at the outcome of already-reviewed items, etc... Just let me know. – Shog9 Oct 19 '13 at 20:28
  • 3
    @Shog9: Can you get a count of how many are "tag:C++" AND "reason:Off-topic-must-have-minimal-understanding"? (Or point me to a parametrized query?) – Kerrek SB Oct 19 '13 at 20:33
  • 1
    If you want to try out stuff about when questions do get closed (not about the close queue itself) in the data explorer, see How to determine why a question was closed from the data dump/data explorer – Gilles 'SO- stop being evil' Oct 19 '13 at 20:36
  • 4
    Do you want fries with that too? – Robert Harvey Oct 19 '13 at 20:36
  • 2
    I could go for some french-fry soup right now... – Shog9 Oct 19 '13 at 20:40
  • You may or may not find this helpful, but here's a query that gets open questions with close votes, sorted by most votes including a clickable link to question (which I made for another reason). Feel free to try to make something useful of it. – Dukeling Oct 19 '13 at 20:41
  • The answer to that is 301, @Kerrek – Shog9 Oct 19 '13 at 20:43
  • @Shog9: Neat, thanks - on the one hand, that's a quantity I feel comfortable tackling; on the other hand, it'll make absolutely no difference to the first n decimal places of the queue size :-S – Kerrek SB Oct 19 '13 at 20:46
  • 1
    Not to give too much away (since, as I said, I'm writing another answer on this that goes into much more detail), but... There is a massively long tail to the queue. I recommend reviewing to improve your own little corner of the site; SO is too big to drink from the firehose @Kerrek. – Shog9 Oct 19 '13 at 20:49
  • 1
    @Shog9 Yet the close vote queue is one giant firehose... – Dukeling Oct 19 '13 at 20:51
  • Not if you filter it, @Dukeling... – Shog9 Oct 19 '13 at 20:52
  • @Dukeling What about filtering by tag? – Frank Oct 19 '13 at 20:57
  • 1
    You can filter by reason and tag, @Dukeling. Try... OT + regex and you'll be looking at under 300 (unfortunately, you can't do AND tag combos, so you can't further reducing this by adding java). – Shog9 Oct 19 '13 at 21:00
  • 2
    I'd want to see, for each language tag: the ratio of questions in the queue to close-voters with expertise (say, a bronze badge in the tag). SO could recruit more mods from areas in which the ratio is high; and concerned folks could start meta or chat conversations to come up with new norms for questions in those areas. – Frank Oct 19 '13 at 21:00
  • 1
    I didn't think through the Bounty thing. Perhaps multiple ones (is that possible?). @Dukeling can those results from your queries be updated? The count one gives a better guess at how many activities may remain to get the whole queue closed. rather than just 85k by some number. – Bill Woodger Oct 19 '13 at 23:20