By Programmers For Programmers

Motivation: community-driven Q&A internet forum has been built by programmers for programmers "with the ultimate intent of collectively increasing the sum total of good programming knowledge in the world". From its launch in 2008 it grew up to over 2,400,000 users. And now is the most popular peer knowledge exchange forum. We analyzed the data available and would like to share several curious facts.


Data were taken from the site. Currently everybody can write a query or rerun one of the previously written

Curious facts:


Initially the reputation concept was developed to allow the site to be self-moderating: questions receive multiple answers and the best answer gets to the top by votes. Following the founder Jeff Atwood vision: "Reputation is a (very) rough measurement of how much the Stack Overflow community trusts you. Reputation is never given, it is earned by convincing other Stack Overflow users that you know what you’re talking about". Then it transformed in some kind of currency, when a user can offer extra reputation points (being transferred from his/her account) for an answer on a question. Higher reputation provides extra opportunities.

Or if we skip reputation < 1 cases, the plot will look like:


Though, the site had been designed and in currently located in US, its intention is to serve the world community. Indeed the site has users spread over the world. The distribution of users in the top 10 countries (by the number of registered users):

(based on profiles where users pointed out their location). The plot is normalized having the 10th country value set to 1, in such a way it's easier to see the difference. While the distribution of the site visitors looks like:

The next curious fact would be how different countries contribute into "the sum of good programming knowledge"

And how much it differ if we will consider only top users with reputation over 1000.

Or top 10 countries by average reputation (reputation > 1)

vs top 10 countries by average reputation (reputation > 1000)

The distributions of reputation by country doesn't follow the same pattern as number of registered users. Though, with some exceptions it's fairly close.