ActionScript and Java appear to be high demand programming languages, thanks to their role in enterprise software and in cloud computing management. As a result, IT professionals with these skills are pulling down amazing salaries. But what sort of cross-section of the industry can developers expect from what languages they write in?

Ben Podgursky, a Software Engineer at Liveramp, posted a blog article where he shared some interesting figure on average household income based on programming skills–specifically language use. His report listing is on 20+ of the most common programming languages alongside with the average yearly income for developers in those languages.

He based his statistics on Git commit metadata plus the Rapleaf API to build and aggregate demographic profiles for popular GitHub organizations. He also used different demographics per programming language instead of per organization. To focus on context and repository, he used GitHub’s estimate of a repository’s language composition, and incomes for all developers who had contributed to a project in at least 50% that language. He then filtered for languages with more than 100 available income data points.

Based on the calculation, ActionScript yields highest average household income of $108,119.47, followed by XSLT ($106,199.19), Java ($103,179.39), Groovy ($102,650.86), Objective-C ($101,801.60) and ColdFusion ($101,536.70). Puppet ($87,589.29) and Haskell ($89,973.82) rank lower than other programming language in the list. Haskell is widely used in academic and academia is not known for good salaries.

Podgursky noted his data is based on open-source projects, which may not accurately represent compensation among closed-source developers. In addition, there are many languages, which are not recognized by GitHub (SQL, among others) and are not listed in the stats.

Misleading calculation?

It is important to note that as people are using Open Source, they’re coming to GitHub to get things like jQuery, MooTools, Erlang Operating System and other languages hosted there. One of the problems here is that self-reporting on household income are notoriously unreliable since the people whose salaries he listed there have just contributed to projects with more than 50% of a given language. The weighting also doesn’t follow certain traditional expectations such as the survey was done with small sample sizes per language, project contributors not actually getting paid for the project in question and companies unwilling to host their code on third party servers.

Many comments on Hacker News and reddit at least reasonably pointed out problems with the data presented.

One of the developers said on Reddit that a lot of people use a wide array of different languages on the job. Saying that someone who does “CSS” makes “X amount of money” is an inherently misleading statistic. Another note is that salaries are largely based off of negotiation, experience, etc. You could have the best coder in the world writing something on the bottom of the salary food chain but he isn’t getting paid as much as someone writing PHP.