Programming languages are just tools. Pick one (Python is a good start for someone from a non-statistical background with little programming experience) and learn the fundamentals, but your main focus in learning how to data science should be more on the methodologies. Languages, frameworks, APIs all come and go and are updated frequently, but what you will end up using them to do and why is not likely to go through such dramatic shifts.
Learn about ethics in data science. Get your statistics solid. Learn how to make and read visualizations; even exploratory visualizations meant only for your eyes can be done incorrectly. Depending on what in data science you want to do, some basic calculus will come in useful.
There are plenty of tools that already do all the heavy mathematical lifting, but you need to know how to use them properly as well as answer if you are asked how you came to the conclusions that you do. I like Khan Academy for the math. It’s slow moving, but I’ve never finished a section with unanswered questions.