How long does it take to become a sports scientist?

What does it take for a mathematics and computer science student to become a data scientist?

• I am an undergrad student majoring in computer science and math. I have completed the core undergrad math and CS courses, passed 4 actuarial exams and interned as a software engineer. What else should I learn in each of the following categories to become a data scientist? --Theoretical math (eg  probability theory, measure theory etc.) --Applied math/statistics (PDE etc) --Theoretical computer science (theory of computation, programming language theory etc.) --Applied CS and software engineering (Design patterns, Distributed systems...) --Other (anything else that does not fall into the above four categories)

• Answer:

  I studied mathematics and computer science and I'm currently working as a data scientist, here at Quora. Some of the most useful classes I have taken in preparation for this role have been statistics classes. Specifically ones which take you beyond the textbook modeling scenarios of first year - i.e. linear models. If there's a good class in non-parametric statistics, and a solid course in bayesian statistics then I think this will cover some useful material. Similarly if there's a machine learning course, this will be invaluable. Time series analysis is a really hot topic right now, as most data analyses have a temporal component, i.e. they are time series, and if you can incorporate this into your experiment or model then you'll have an edge in the job market. For some pragmatic courses on the engineering side, I think a class in large-scale databases would frequently come in useful. Or just any database class that you can take, provided that it's not taught by an awful professor. I never did this and frequently regret it. With regards to measure theory, your chance of using it within a professional setting is incredibly slim. However it will help you understand probability and thus statistics on a more concrete level. Algorithms classes translate straight into industry and I think are really useful - a fairly large chunk of CS theory (I'm thinking the contents of Sipser's ToC) is less useful on a day to day basis. Having said this I'm quite a big believer in letting your interests shape what you study. If you think a topic is interesting and you're passionate about it, you will probably absorb a lot more of the content. I took a course in information theory; the concepts that Shannon thought up over 60 years ago pop up randomly from time to time and although it's hard to quantify its value in a professional setting - it seems like a no-brainer in retrospect. If I didn't find it fascinating however, it probably would have been a waste. Similar sentiments for classes like network analysis, graph theory, quantum information theory, measure theoretic probability, group theory and number theory which have captured my imagination and enthusiasm on many occasions. You chose well with maths/compsci - it opens up so many doors. Good luck!