I was checking out this year's #StackOverflow annual developer report that few things caught my eye:

1. They have deliberately drawn the wrong conclusion about most popular operating system. They have interpreted #Linux distro as operating system, so they have separately provided numbers for Ubuntu, Arch, Fedora, and etc., and have concluded that "Windows is the most popular operating system for developers, across both personal and professional use".

https://survey.stackoverflow.co/2024/technology#1-operating-system

1/🧵

@yo
The issue is that they are not providing the raw (or cleaned) data to be analyzed "correctly". I think there are a lot of statistical assumptions that are not clarified in the methodology. Also lack of reporting the variance makes it pretty unreliable about trends.

For example the report says 0.6% of participants have primary and elementary school education!

Perhaps the only "reliable" part is the age-range of users:

https://survey.stackoverflow.co/2024/developer-profile#6-age

ALT

@Mehrad Oh, they don't? I think they used to do this in previous editions.

I've always thought there's a couple of interesting data analysis things to do there. (E.g. show stuff which is different when you fix one variable; for example, how devs from country X differ from the average.)

Shame.

@yo
It might be there, but at least I didn't find it. I would be happy if someone can point me to the direction of finding the data for proper analysis.