Without more information from the OP, this answer will be a bit tentative. I'm not sure that a low GPA is your biggest issue, though some schools have rules that they enforce quite rigidly.
I think the biggest obstacle for you now, that you can correct somewhat at least with time and effort, is that your academic background seems to be haphazard rather than focused. Normally a master's is quite focused and depends on the student having certain skills, usually from their undergraduate years. You don't seem to obviously have that.
Learning Python, of itself, will help you very little, actually as it is just a tool. But you may be missing most of the content of the undergraduate education, including, perhaps the vital Algorithms and Data Structures courses, as well as Database and a few others.
Physics (at most levels) is an actual science. That seems to be your background. Computer Science isn't a science at all (in most areas) and hasn't been for half a century. Data science is a bit between, however, as it can be experimental, not just constructive. In CS, we mostly build things, rather than search for the truths provided by the cosmos.
It might be, that if you want to study big data in Physics specifically, that you might find a program somewhere that focuses on that, though it would likely be in a Physics department. If you can find that, your background might be seen as helpful.
But in general, if you want a MS (MSc) degree you should spend some time (not a small amount of time) focusing your studies in one area primarily that will help you be seen as having the needed skills.
While broad interests in life are valued and valuable, academia mostly values specialization. Studying data science online may help you provided you do enough of it and do it well enough, but your broad background may only make it harder. If you can find a way to get some of the background in an actual educational institution (with a mentor) it would (for most people) be better than online courses, which are harder to judge no matter how good they are. There is too much variability in student outcomes to trust them completely. Focus.