Unfortunately those search queries are worse than "ballpark" figures as they're mostly based on keyword matching, not the amount of code the developer is writin

Unfortunately those search queries are worse than "ballpark" figures as they're mostly based on keyword matching, not the amount of code the developer is writing in that language for that job. There are certain search queries that will always hit high. Of course, JavaScript is going to be in almost any job posting that's web-related, even if it's not true hardcore JS coding. Perl is used heavily for development automation, so it's also going to have a big showing, even if it isn't the core language. You'll also tend to see lines like "Previous scripting experience with Perl, Python, Ruby a plus" for Java postings. If it could be monetized, doing more sophisticated data mining into job postings for actual popularity trends would be awesome, especially with a decent granularity. It wouldn't be difficult to train a supervised ML algorithm with a set of keyword-tagged job postings with weights as to how significant a certain set of skills would be used at a job for a certain job posting.