I'm currently pursuing a bachelor's degree in physics from a university in the UK. Most data science jobs here have a strong preference for people with PhDs in numerate degrees like physics and maths. I don't understand the point of spending 5 years of my life doing research in something I don't want to pursue a career in.

I'm confused about what master's degree I should choose if I want to apply for a job straight after graduating. A physics degree kind of leaves me hanging in the middle in terms of skills. I have done some analysis for experimental data using Python, but only very basic stuff. A bit of predictive modelling and C++, lots of maths, but barely any statistics. So I could either get a degree in CS to get better at programming or get one in statistics. Which one would give me the skills most relevant to data science?

Also, would you recommend that I get a diploma/certification in machine learning?

  • $\begingroup$ Now that you are a student, get a computer science or statistics course or audit one. After getting a sense of any of those, you can decide. You can try MOOCs as well. you can take a course for free and start learning as soon as possible. You can decide after going into data science. $\endgroup$
    – Hamideh
    Commented May 11, 2016 at 15:07

1 Answer 1


I think the best thing for a beginner is some practice to let the theory sink in. (If you need book suggestions, do a search.) If you are weak on computer science, contribute to an open source project like scikit-learn. If you are weak on data analysis, compete in a Kaggle competition. Write a paper and/or a few blog posts to prove your skills. Do this now while you are in school. An undergraduate physics degree should be enough to get your foot in the door if you do the above. A degree will rarely get you a job by itself; you still have to pass the job interview, and that's why I recommend getting practice.