The real question is not how to visualize the data (or any of the conclusions)... its how the author cleaned the data. This particular data set is fraught with input errors, intensional mis-direction, and cultural bias.<p>A simple example that has major implications are the number of full time people that put in < $15K a year salaries. Which in the real world is not legally feasible in the US or any western country. Further analysis leads to input errors, and cultural differences (Like stating monthly salary versus, annual salary... or Net versus take-home pay).<p>I would be interested in how they handled these aspects and what the author thinks about the historical data from the stack overflow surveys.