A majority of my work is glueing things together (aka 'data science') with the occasional big application or number cruncher thrown in there. For an unexplainable reason, I don't like python. However, it's exceedingly difficult to justify anything other than python these days.
Interesting ... I use Python a lot and like it. But have been learning R to come to grips with data science / machine learning material. I believe that some people are using Julia.
I would suggest use the language(s) that you are most productive with and give you the libraries and tools to get the job done.
I mean, that is an unassailable point. I guess I would say that I feel productive in most languages after the learning curve. I get caught in the cycle of "oooo what if I'm more productive in x, lisp, or z"
Not that difficult: I use Cython because it's incomparably faster than Python, and anything Pythony I don't like, I can write in C-style, and vice versa.
I would suggest use the language(s) that you are most productive with and give you the libraries and tools to get the job done.