If you can do meaningful data science in one language, and nothing else, you’re ahead of a whole bunch of curves. Get solidly good at one thing.
You’ll pick up a bit here and there about other techs in the course of the job. That will a) help you pick the highest leverage next thing to learn, and b) anchor you when you start learning them.
Stick to a theme. You don’t need to know web sockets to do data science.
For followup learning, pick complementary technologies (instead of Python and R, which compete; or R and and Docker, which are virtually unrelated; try either Python plus tensorflow or R plus tensorflow — because either of those pairings can be used together to increase your capabilities.)