Masters in Statistics vs in data science? Is DS just buzz?

[annoyinggirl]

which do you think is the smarter choice, in terms of employ-ability: ms in statistics or Ms in data science? do you think "data science" is just a buzz words that will die out? is a data scientist someone who can't program as well as the computer engineer, and can't build models as well as a statistician - a jerk of all trades, master of none? And is a jack of all trades a favorable trait to have in data analysis, esp in the age of computers? Or do you think people will realize that it is best to have statisticians build the models and then have programmers do the programming? But are not most statistics models just overkill for industry? programming skills are more important today in data analysis?

 

[Locrian]

I'd advise looking past the words/labels and looking at the underlying skills. The words "statistics" and "data science" both encompass a lot of things, and probably more than they should.

I do think everyone needs to know how to code, even if it's in a fairly high level language.

 

[Stephen Tashi]

Or do you think people will realize that it is best to have statisticians build the models and then have programmers do the programming?

That isn't necessarily the best organization. In fact, only in a very large or a very bureaucratic organization would there be designated programmers who only have the responsibility of programming.

That organization might have been common in the old days - when it was also best for the technical staff to write their reports in long hand and for the secretarial pool to type them.

A situation where I have seen programmers who only program is when a contractor is executing a contract with a large organization (e.g. a US government organization) and the contract specifically states that the contractor is only responsible for providing programming services. In that case, programmers for the contractor can refuse to do analytical work and demand detailed specifications for what they are supposed to implement.

I've met programmers who feel this is only "professional" way to operate. Years ago, I knew analyists who felt it was beneath their dignity to program. However, nowadays I think it's best is to have employees who can tackle a variety of jobs - analyze, program, pack boxes when the firm moves to a new office, etc.

 

[StatGuy2000]

I would agree with Locrian and Stephen Tashi that it's important to look past the labels and focus more on the specific skills you gain in either program.

What I can tell you is that most MS students in statistics (including myself) know how to program (as well they should), and certainly during my career have programmed and built statistical models in a variety of different career areas (currently working in the pharma/biotech sector). The MS program in statistics should provide an option for students to take computer-intensive and applied courses in addition to the more mathematical courses.

My concern about the data science MS program is that the degree is still relatively new, and the quality of the data science MS programs may be highly variable. And at the end of the day, I'm not sure you'll necessarily end up with a skill set in a data science MS program that is any different from a MS in statistics. I would look very carefully at what the requirements of either program cover, see if internship opportunities are available, etc.

 

[MarneMath]

Any Statistics program worth it's salt will require their students to program. While it may not be advance programming, the truth is one doesn't need to know many advance concepts in computer science for statistical programming. I think StatGuy concerns are well laid out. In my experience, Data Science MS from even well known schools leave a lot to be desired. I think they make exceptional Business analyst or traditional BI people, but with regards to what I consider Data Science, there isn't enough sufficient mathematics in those programs to leave me feeling confident in their ability to interpret the intercept of a regression slope properly.