8
$\begingroup$

I've just started playing about with SEDE (a very nice tutorial here) and one of the queries I eventually wrote was this one on the votes per month. They had quite a distinct trend I hadn't really expected. Thought someone might find this interesting.

Votes per month on WB

The months are numbered as you would expect (1 being January and that low point at 5 being may).

I know it isn't 'activity' exactly and I might try to change that (take into account number of questions/answers made maybe) but still a measure of something all the same.

EDIT:

After Whoosh's comment I thought I would see how age was distributed across WB accounts. I did that (I think) in this query. Which produced the graph below: WB Age Dist

This seems to support that a lot of WB users are under 25 but the bulk are still over. Of course some people have made up their age (unless we have a three year old and 42 97 year olds on here) but I would expect a trend to still show.

I'm going to try to make a query to pick out the votes cast and posts made by each age (it could be the younger users ask or answer more questions or vote more liberally and that leads to the trend we see).

$\endgroup$
13
  • 1
    $\begingroup$ Interesting. That tracks with what 25k users can see in the site analytics. It'd be interesting to see a query that takes into account the number of questions and answers, as you suggested. It might also be interesting to take into account the number of active users, if you can figure out some measure of activity. You can see total upvotes and downvotes cast by a user, but not over time.) $\endgroup$
    – Monica Cellio
    Commented Feb 27, 2017 at 17:12
  • $\begingroup$ Could the dip be related to the fact that we've been a site for 2.5 years? That is, once we've gotten to the 3 year mark, every month will have happened three times? $\endgroup$
    – DaaaahWhoosh
    Commented Feb 27, 2017 at 18:00
  • $\begingroup$ That could well be part of it. Do you know which the site started? $\endgroup$
    – Lio Elbammalf
    Commented Feb 27, 2017 at 18:07
  • $\begingroup$ In relation to my previous comment, I've made this query which only uses data from the last year, the numbers are much more similar. It looks like we started in September 2014, that's when the oldest question was asked. $\endgroup$
    – DaaaahWhoosh
    Commented Feb 27, 2017 at 18:09
  • $\begingroup$ Oh, wait, I think I did something wrong. After rechecking I appear to now be getting a similar trend... $\endgroup$
    – DaaaahWhoosh
    Commented Feb 27, 2017 at 20:17
  • 4
    $\begingroup$ Too many damn 20-year-olds taking time off in May for honeymooning! Get back to world buildin', not world populatin', you sex-crazed kids! $\endgroup$
    – SRM
    Commented Mar 2, 2017 at 3:25
  • $\begingroup$ I think i got almost right the post by age query if someone want to review it data.stackexchange.com/worldbuilding/query/634941#graph (i had to remove the 0 yo to have a sensible graph however) $\endgroup$
    – Sefa
    Commented Mar 2, 2017 at 9:54
  • 2
    $\begingroup$ And BTW, the 3 years old is community $\endgroup$
    – Sefa
    Commented Mar 2, 2017 at 10:04
  • $\begingroup$ @MonicaCellio If you are fine with considering lurker as active users we can always use the LastAccessDate as a mesure of activity. Else it's still possible to go look for the date of the last post/comment/vote date, but that make a complexer query (and so more chance of bug/false positive) $\endgroup$
    – Sefa
    Commented Mar 2, 2017 at 10:34
  • $\begingroup$ @Sefa Nice query, seems to be popping up with quite a few sharp peaks - either wb appeals to very particular ages or there are some particularly active users. I used LastAccessDate as my limit - assuming there weren't too many lurkers however I'm not quite sure how it calculates it (does it ignore if they have been on other SE sites but not WB?) $\endgroup$
    – Lio Elbammalf
    Commented Mar 2, 2017 at 11:51
  • $\begingroup$ @LioElbammalf LastAccessDate is segregated by site (or at least i dont have the same here on WB and PPCG) $\endgroup$
    – Sefa
    Commented Mar 2, 2017 at 11:56
  • $\begingroup$ a few more query related to ages and votes this time raw stats here and ratio here $\endgroup$
    – Sefa
    Commented Mar 2, 2017 at 15:44
  • $\begingroup$ @Sefa Interesting. That doesn't show a particularly substantial number of the votes concentrated below early twenties - Some but still, is it enough to support school being the reason for the dip? $\endgroup$
    – Lio Elbammalf
    Commented Mar 2, 2017 at 17:39

1 Answer 1

Reset to default
3
$\begingroup$

I initially thought that this was due to the site being 2.5 years old, as the months that have not been repeated 3 times were the ones with the least activity. However, after modifying the query to look over the past two years (thus every month should have been visited twice), and even when looking at questions/answers, it appears the same trend exists:

link to query image of query results

Funnily enough, this trend doesn't seem to exist for other sites (part of my confusion was due to the fact that I was looking at stats for Stack Overflow for a while).

The only explanation I can come up with is school. People are really active in the summer, but less so as the months go on, then activity picks up again around winter break, before starting a long downward slope for the spring semester and final exams.

$\endgroup$
3
  • 2
    $\begingroup$ To explore your idea of it being school related I plugged the query into a couple of other sites which are specifically school related (hoping to see the inverse trend that when people are most active in school they need help on questions). Physics SE does show this a little, a peak in Feb, April, drop off throughout the summer and coming back to peak in October The summer and winter dips are noticeable in mathematics SE too. Neither are quite so drastic as on WB but I suppose this does indicate a younger audience on here that may draw a little from all other subjects. $\endgroup$
    – Lio Elbammalf
    Commented Feb 28, 2017 at 8:24
  • $\begingroup$ I made another query (in edit above) to see what the distribution of ages was like among WB users and whilst we do have a substantial number at school age it isn't as large as I would expect for the trend we're seeing. I've still got a couple of query ideas though (as stated in the edit). $\endgroup$
    – Lio Elbammalf
    Commented Mar 1, 2017 at 11:46
  • $\begingroup$ Perhaps since WorldBuilding is decidedly less "academic" than the likes of Physics, Mathematics, etc., as the end of the term approaches and the work-load increases, students begin focusing more on their core subjects so they can pass their exams. $\endgroup$
    – user26892
    Commented Mar 13, 2017 at 14:45

You must log in to answer this question.

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