Can only speak from my experience. I learnt to program in C++ as part of a Physics degree and then went into research. At the time, people mostly used MATLAB o

Can only speak from my experience. I learnt to program in C++ as part of a Physics degree and then went into research. At the time, people mostly used MATLAB or Mathematica for interactive stuff and C/C++/Fortran for higher performance bits. Mathematica was never paid for by the University while I was doing my PhD which meant I had to pay for it out of pocket, and MATLAB was clunky to do anything not matrix based. I started using Python just after version 2.7 came out because someone told me to download Enthought Canopy (a precursor distribution of scientific libraries for Python, like Anaconda is now) , "Hey, you can use Matplotlib and Numpy and get 90% of the way there by barely modifying your existing code". Then I came across things like FEniCs (finite element simulation framework) which didn't have a real equivalent in MATLAB/Mathematica world and which C++ world tools were difficult to use and tended to require a custom programming language. After leaving research I've worked as a developer and keep coming back to Python. Currently I work in web and use Django in my day job. I'm super comfortable and productive in the language.