2
$\begingroup$

I was wondering if any of you was computing statistics about the site. Stuff like:

  • percentage of answered questions
  • questions rate
  • trends with time (for example, I guess during this time of the year the activity will be lower!)
  • etc.

UPDATE I found this questions

Statistics on Cross Validated usage

and I started using SEDE to get some stats, but I was wondering if someone else already did this job, or was interested in collaborating on it. As a start, my hypothesis that Q & A drop down in December (with As dropping more, in percentage, than Qs) seems plausible:

enter image description here

$\endgroup$
9
  • 2
    $\begingroup$ A variety of time series, presented in graphical form, are available to moderators--but we are explicitly not allowed to share them. There is a strong seasonal decline in questions and answers at the end of each December and beginning of each May, clearly associated with academic calendars. You can obtain comparable statistics through the SE API. $\endgroup$
    – whuber Mod
    Dec 23 '17 at 17:43
  • 1
    $\begingroup$ Found it: sese.evbpc.com/Sites/…. $\endgroup$ Dec 23 '17 at 19:00
  • $\begingroup$ stats.stackexchange.com/users/8373 $\endgroup$
    – usεr11852
    Dec 23 '17 at 20:37
  • $\begingroup$ @AndreSilva your first link is great! I would accept it as an answer $\endgroup$
    – DeltaIV
    Dec 23 '17 at 21:22
  • $\begingroup$ @whuber thanks for the tip. Looking at the time series I would have actually hypothesized that the summer dip was in July, but I can be wrong. $\endgroup$
    – DeltaIV
    Dec 23 '17 at 21:23
  • $\begingroup$ @usεr11852 David Robinson is a nice guy but I'm not sure how his profile could help me. As far as I know, he only blogs about Stack Overflow statistics, not Cross Validated ones, and I can't contact him on CV...maybe you were suggesting to get in touch by mail/Twitter? $\endgroup$
    – DeltaIV
    Dec 23 '17 at 21:26
  • $\begingroup$ I believe the best action would be to make this Q a dupe, and add that link to Glen_b’ s answer or post it as a new one there (I prefer the former). But feel free to use it in a self-answer as well. Tks. $\endgroup$ Dec 23 '17 at 21:38
  • 1
    $\begingroup$ @DeltaIV: It was just for a jokey direct answer to your question "Does anyone compute statistics about Cross Validated?" (In retrospect I should have put a smiley alongside it :) ) $\endgroup$
    – usεr11852
    Dec 23 '17 at 21:52
  • $\begingroup$ @AndreSilva I don't think it's exactly a duplicate, though I did explain myself really badly, because I was more interested in 1) existing analyses of Cross Validated data (for example, a blog post/web app/whatever from one of you guys), which discussed the data and drew conclusions, or 2) in collaborators to perform such an analysis. I knew about SEDE, but that doesn't even make plots, it allows me to extract data which I then need to analyze. I'm not sure if I should edit my question, because it seems clear now that no one has written such a blog post. $\endgroup$
    – DeltaIV
    Dec 23 '17 at 22:02
3
$\begingroup$

The SE team tracks a lot of statistics like these. You can find some of these statistics on the SE sites page. Specific pieces of information are also computable with specialized queries in the SE data explorer, although someone either needs to have already written the query or you need a bit of SQL sophistication to get them (they do have a quick tutorial). As @whuber notes, there are some timeseries that are already prepared for convenience; they are available to moderators and to users with privileges for reputation >25k.

$\endgroup$

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