10
$\begingroup$

There are every now and then discussions about where the Worldbuilders are active outside WB.SE. The definition of active is of course hard to define, but a user reaching a certain reputation usually indicates a certain activity (even if that activity could be quite old). In any case, I wrote a SEDE query:

Cross-site Overlap

It requires to set a minimum threshold. Find the list of all SE sites and compare on each site how many users have more than the given reputation threshold on both WB and the other sites from the list. Anyone can try it out.

But just to give some example, if we follow the standard of SE that requires 200 reputation to be considered active, we get the following overlap for WB:

  Stack         |   User overlap
------------------------------------
  StackOverflow |  53.3 %
  SFF           |  26.4 %
  ELU           |  23.0 %
  Programmers   |  16.4 %
  SuperUser     |  15.7 %
  The Workplace |  15.4 %
  RPG           |  15.0 %
  Meta          |  14.8 %
  Security      |  12.8 %
  Gaming        |  12.4 %
  Physics       |  11.7 %
  Academia      |  11.2 %
  Puzzling      |  11.0 % 
  Code Golf     |  11.0 %
  Maths         |  11.0 %

for the top 15 overlap.

If you are curious about who are the users present on two concrete sites, you can use that other query.

Please note that I am a total beginner in SQL, so don't be too harsh about non-optimum syntax.

||||||
$\endgroup$
  • $\begingroup$ Note that the average rep of our users is less than 200, which means the majority of people on WB aren't active here. $\endgroup$ – Frostfyre Mar 16 '16 at 14:26
  • $\begingroup$ @Frostfyre while that's true, that's the same on most sites. SO's average is 123, and 172 on SFF. It makes sense many people come ask a question, got their answer and look somewhere else. If you lower the threshold on the query from above, you'll see a stronger effect of the other sites user base: 92.4% of WB users have more than 100 rep on both WB and SO. $\endgroup$ – clem steredenn Mar 16 '16 at 20:11
  • $\begingroup$ Which leads to the question: Are there really that many developers on WB, and why? $\endgroup$ – Frostfyre Mar 16 '16 at 20:14
  • 1
    $\begingroup$ Green presented a view of that on chat yesterday. It's possibly related to the reason why people joined SE. Many first got to learn about it through SO. And then arrived here. $\endgroup$ – clem steredenn Mar 16 '16 at 20:16
  • $\begingroup$ @bilbo_pingouin I totally second you on that. I learned that Stack Exchange actually existed when I used SO while I was programming. I saw some Worldbuilding questions in the hot questions bar which got me interested, and that's how I learned about WB, because we are generally a niche community. $\endgroup$ – fi12 Mar 17 '16 at 0:51
  • $\begingroup$ @fi12 also how I got here. I expect most people do it that way. $\endgroup$ – Separatrix Mar 28 '16 at 7:34
  • $\begingroup$ Someone with a reputation of 101 here may have simply joined to vote on a specific post. Many users qualify for the +100 reputation on all registrations. Someone with 200 or more actually participated on the site with at least one positively received post. $\endgroup$ – Brythan Nov 5 '16 at 11:34
9
$\begingroup$

It's been something like seven months, and I was curious to see if things have changed at all. It turns out they have.

I used the same query with the same minimum rep requirements. Here's what I got, for the top 15 (to one decimal place, rounding):

  Stack         |   March User Overlap   |   October User Overlap   |   Change
-------------------------------------------------------------------------------
  StackOverflow |  53.3%                 |  49.8%                   |   -3.5%
  SFF           |  26.4%                 |  24.7%                   |   -1.7%
  ELU           |  23.0%                 |  21.0%                   |   -2.0%
  Programmers   |  16.4%                 |  14.3%                   |   -2.1%
  SuperUser     |  15.7%                 |  14.0%                   |   -1.7%
  The Workplace |  15.4%                 |  14.9%                   |   -0.5%
  RPG           |  15.0%                 |  14.0%                   |   -1.0%
  Meta          |  14.8%                 |  12.7%                   |   -2.1%
  Security      |  12.8%                 |  12.7%                   |   -0.1%
  Gaming        |  12.4%                 |  12.7%                   |   +0.3%
  Physics       |  11.7%                 |  11.1%                   |   -0.6%
  Academia      |  11.2%                 |  10.5%                   |   -0.7%
  Puzzling      |  11.0%                 |  11.9%                   |   +0.9%
  Code Golf     |  11.0%                 |  Not in Top 15 (9.0%)    |   -2.0%
  Maths         |  11.0%                 |  10.3%                   |   -0.7%
-------------------------------------------------------------------------------
  Travel        |  Not in Top 15         |  10.5%                   |     ?

A couple sites switched places - Code Golf dropped to 16, while Travel rose into the Top 15, and The Workplace and Puzzling saw increases in position. On the whole though - and this is the amazing thing - most changes were negative.

This could mean a few things - maybe more people are active on smaller beta sites, and are spread out, but I doubt it - but the conclusion I draw is that we're getting more people who are on Stack Exchange just for Worldbuilding Stack Exchange, and are participating enough to get a decent amount of rep. That's awesome. It means that we're finally attracting people from outside Stack Exchange, rather than just curious passersby who treat us as a novelty.

Is this a small increase? I don't know; the chart shows fluctuations of up to 3.5%, and even 0.5% is non-negligible. But I think it's a good sign that the site is growing. I might run the query again this March to see if we continue to see drops in the percent of users active on other sites, but I think the results I got today are still accurate.

||||||
$\endgroup$
  • $\begingroup$ I'd be curious how it evolved since last year... $\endgroup$ – clem steredenn Oct 24 '17 at 8:22
5
$\begingroup$

This produces interesting results, but I'm not sure that people understand what this means. What this tells us is how many of the Worldbuilding users come from each other site. By its nature, this query tends to produce the highest results from sites with lots of users. If we flip it around and ask how many users on other sites participate on Worldbuilding, we get very different results:

22.41 StackExchange.Hardwarerecs
22.16 StackExchange.Politics
21.74 StackExchange.Opensource
21.43 StackExchange.Ai
20.65 StackExchange.Mythology
20.42 StackExchange.Law
20.42 StackExchange.Space
19.13 StackExchange.Moderators
18.64 StackExchange.Crafts
18.15 StackExchange.Astronomy
17.79 StackExchange.Lifehacks
16.55 StackExchange.Health
16.52 StackExchange.Engineering
16.33 StackExchange.3dprinting
15.64 StackExchange.History

Cross site overlap; based on other site

What this query tells us is how many of the other site's active users are active on this site (the original query tells how many of our active users are active on another site).

I like this query better because it pushes Stack Overflow down towards the bottom (few of their active users are active here while many of our users are active there). SF&F moves down to the middle, somewhat behind Writers.

||||||
$\endgroup$
  • $\begingroup$ I think this is a nice complement. However, it has the inverse bias. This one favours the sites with lower number of users. To take the case of SO, yeah, few of SO members are active here. But most of our members have an account there. And many somewhat active. But I think the two of them show a nice picture of the full overlap. $\endgroup$ – clem steredenn Nov 5 '16 at 12:45

You must log in to answer this question.

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