Stack OverflowJanuary 24, 2010 at 10:51 am | Posted in Programming | 3 Comments
I only recently discovered the site Stack Overflow and I was immediately fascinated by the concept. For those who don’t know Stack Overflow : it is a programming Q&A site. The nice thing is that you can earn points when you put a relevant question or answer on the site. This is a very rewarding concept and got me addicted within a few minutes. After one week of use I now have 96 reputation points and even 4 badges. Reminds me of my time back at school.
Now 96 points isn’t very much, especially not compared to the number one John Skeet who has collected a whopping 133k reputation points. However, this made me curious about how the reputation was distributed over the number of users. The site has about 125k users. You can get a paged overview of those users with each page containing 35 users. I decided to plot the average reputation against the page number and came up with the following graph:
What surprised me is how neat this graph looks. A very smooth long tail distribution. With those few questions I’ve answered I’m already on page 880, which is within the top 25 %. To get into the top 10 % (around page 300) is still relatively easy. After that the curve increases sharply, which is of course as expected.
Next step is to do a bit more analysis on those numbers. I have to figure out if Stack Overflow has a public API. Would be very cool if I could combine that with my new favorite language Clojure. Another thing that interests me is on how to use this concept on performance reviews. Wouldn’t it be great if your boss would rate your performance on your reputation? That would motivate people to share their knowledge even more with their co-workers. More on that in another blog.