Anyone got a good practice to become a data scientist?

Is that when I get Bunsen burners out and light USB drives on fire?

One step at a time... First become data, then go for the scientist part.

Data science is such a flexible term. Probably because it's really just a buzz word used to describe someone who performs statistical analysis on large amounts of data, and performs some basic machine learning. So step by step:
Learn calculus statistics
Learn linear algebra
Learn python,
Tensorflow
Scikit learn
Pandas
Numpy
Learn R
Finish your CS degree
Get an internship
Eventually transition to a real job

Or if you have no CS degree, spend years and thousands on coding boot camps that will never get you a job. Or just get a degree in CS.

As a note, I don't say this to be an ass. I say this as someone with a masters degree, specialized in machine learning and statistics. It's hard out there. I still can't get a job as a data scientist, because everyone wants really experienced players in that game.