i've heard it characterised something like this: you need to program better than the average statistician, and apply statistics better than the average programmer.
I think the bar is a LITTLE higher now. If I could add one skill, its being able to read math language. It's a great skill to be able to read about a new technique on arxiv or a blog and be able to understand it beyond the surface level.
In fact being mathematically literate is really the bee's knees. Its why I suggest anyone in a stem undergrad to take their math department's Real Analysis or (my favorite) Abstract Algebra. The content is rarely relevant any jobs, but then you'll be able to teach yourself almost anything in the future.
If you are serious about this question, you should begin by specifying what "data scientist" means in this context. Otherwise, whatever answers your query engenders will be lost in the noise.
I have read that it is much more important to have a strong mathematical background than a strong programming background to really be good at data science. Many data science positions in the world are looking for individuals who specialized in statistics and related math fields in school.
9 comments
[ 2.9 ms ] story [ 35.4 ms ] threadIn fact being mathematically literate is really the bee's knees. Its why I suggest anyone in a stem undergrad to take their math department's Real Analysis or (my favorite) Abstract Algebra. The content is rarely relevant any jobs, but then you'll be able to teach yourself almost anything in the future.
The former is more math focused, the later is a pretty solid blend of Math and Comp Sci / DB work.
http://blog.udacity.com/2014/11/data-science-job-skills.html
http://www.informationweek.com/big-data/big-data-analytics/3...