In Data Science R and Python are the two big players, when it comes to programming languages.
R has a rich bunch of plugins and a scientifically designed pipe of data processing. The atomic variable is already an array. You get the first dimension for free and the first dimension is the expensive last mile in coding.
Due to its origin R is quite odd in other aspects. Typing an assignment already takes two characters, etc. You also can't really speak of object orientated programming.
Python is clean where R is odd. I seems a lot of the strengths of R have been ported to python in form of libraries. This sounds like Python may be the better choice.
In both cases I miss encapsulation. Missing encapsulation is an issue in community driven development. You always fear to break extensions, while cleaning up the internals of a module or class. It's difficult to split up the development into multiple teams. A core team has to confirm every change, because every change is more or less a breaking change for extensions. The core team again is limited in number due to rules of social dynamic. Missing encapsulation means, the project does not scale well and the core libraries are poorly refactored.
Imagine the good fairy comes and grant's you three wishes for the perfect language. What are your three wishes?