*by People Who Took a Stack Overflow Survey.
Generally, having a random sample is more important than having a large sample size. The fact that lots of developers were surveyed means the survey tells us nothing about the language usage of all developers, since they were self-selected on a particular platform.