How to become a data scientist?

Simply put, the principles and foundations are on the math side. Linear algebra (matrix representations and operations) is the foundation of the foundation, calculus (derivatives, limits); data processing of course requires programming, so C/C++/Python either way, data structures can be learned just to make your programming smoother, but programming is not the core of data processing.

Mid-level courses, probability theory + statistics (a lot of data analysis based on statistical models), linear programming + convex optimization (statistics to the end is still solving an optimization problem, of course, there are pure optimization models do not use statistical models) and then higher-level courses, that is, some of the postgraduate courses, it is more specific, you can look at the project you are doing and then choose to electives. For example: Probabilistic Graphical Models, Nolinear Programming, Integer Programming, Machine Learning (in fact, machine learning, learning are some statistics and optimization), image processing, deep learning, neural networks, and so on. And so on and so forth. If you learn Mid-level, then do a few real projects, you will be able to get started. You'll have to do a PhD in research to get to the higher levels.