It's a nice, informal study. I'll mention this looks at job turnover (and new jobs), not raw numbers of jobs. It's a good proxy for general popularity, but it

It's a nice, informal study. I'll mention this looks at job turnover (and new jobs), not raw numbers of jobs. It's a good proxy for general popularity, but it will have errors (likely 2x or 3x). Java is more popular than Perl, for example, but I don't think we can say how Java, JavaScript, C/C++, and Python compare. Or how Perl and R compare. My experience with Java management culture, for example, is that it lends itself to very high turnover. Programmers are viewed as interchangeable resources, and laid off and rehired at will. JavaScript and Python are used in startups and on newer projects. New projects have high infant mortality rates (which may lead to layoffs at startups, but are likely to be replaced with new projects at bigger firms). C and C++ tend to be things like embedded systems; often incredibly stable jobs. I can't say my observations speak for the industry as a whole, but those do make for errors with this as a proxy. I like to take several studies like this: Hiring, Dice, surveys, etc., and combine into a coherent picture with a clear understanding of error sources and levels. If github, dice, and a survey all point the same way, it's actually a pretty strong indicator.