I'm currently finishing up a B.S. in mathematics and would like to attend graduate school (a master's degree for starters, with the possibility of a subsequent Ph.D.) with an eye toward entering the field of data science. I'm also particularly interested in machine learning.

What are the graduate degree choices that would get me to where I want to go?

Is there a consensus as to whether a graduate degree in applied mathematics, statistics, or computer science would put me in a better position to enter the field of data science?

Thank you all for the help, this is a big choice for me and any input is very much appreciated. Typically I ask my questions on Mathematics Stack Exchange, but I thought asking here would give me a broader and better rounded perspective.

  • 4
    $\begingroup$ You might want to try the data science stack exchange for this Q. $\endgroup$ – Jeremy Miles Nov 20 '14 at 17:14
  • $\begingroup$ @JeremyMiles: That's a great idea, I just recently joined up there so I'll give it a go. $\endgroup$ – user60305 Nov 20 '14 at 17:15

EDIT: Just try to add some more words.

As a PhD student in Biostatistics, I feel great with what @Frank-Harrell said. And that's quite correct!!! Students from our department have great job placements after graduation.

On the other hand, @StasK cited the article "Aren’t We Data Science?", but titled it as "statisticians are not recognized as data scientists". This is somewhat misleading to me. Statisticians might not be titled as data scientists. But who else can formally claim that? Anyway, what the article states, at least to me, is that statistics has the great potential to contribute to data science. The main issue, if there are, that impede Statistics to promote data science is that people from Statistics are not well trained for large-scale computation and efficient programming. Cited from that article is following.

And to statistics. Statistics has enormous potential to contribute to data science. There are open research problems requiring that classical statistical methods in sampling, design, and causal inference be “scaled up” to be feasible with massive data sets. Few of the computer scientists and others who dominate the data science landscape are well-versed in these concepts, and many take an “algorithmic” view of data analysis. Data science needs statistical thinking and new foundational frameworks—for example, what is the “population” when one confronts the Big Data generated by Google?

In fact, many businesses are beginning to collect data prospectively for internal testing and validation, and there is little appreciation for the power of design principles. Statisticians could propel major advances through development of “experimental design for the 21st century”!

One can arguably say that Computer Sciences are at better position but just lacks the statistics thinking. But to me, I regard the two main component as the "brain" and "hands"! If the experiment design is flawed in the very beginning, or if the inference is biased at the very end, we will end up with a totally different story about the conclusion and business strategy.

To put it simple for all I wish to convey here, data science practitioners really need both great statistical thinking and programming.


To decide which degree you are going to pursue, you have to get to know what skill sets that qualify you to work in data science area. Based on what I've known, if you wish to enter the data science field, what "hard" skills you would wish to be equipped are basically twofolds: the strong analytical ability, and good computation and programming skills. You can go to Quora and search like "data science", "data scientist", etc, to get some feelings about what the field looks like, and what you need to prepare for that area. Here are two questions from Quora you might wish to go through:

  1. What is data science?
  2. How do I become a data scientist?

Some questions like that, you get my point.

(The soft skills, like oral and written communications skills and team-work ability are also very important. And in some circumstances, even more important than your analytical skills to some extend. But certainly the discussion on soft skills is certainly off topic for the sake of your questions.)

Now back to your questions.

What are the graduate degree choices that would get me to where I want to go?

Once you have clear vision and deep thinking about what you need to learn, you should be able to answer this by yourself. My suggestion would be Computer Sciences, Applied mathematics or statistics, Biostatistics, Physics, Engineering, or any other degrees that heavily involve analytics and computation. Essentially, an interdisciplinary degree that help you train both data analysis and programming will definitely win you a great position for working in data science area.

Is there a consensus as to whether a graduate degree in applied mathematics or statistics would put me in a better position to enter the field of data science?

I am not aware of whether there is such consensus formally acknowledged by academic researchers or industrial practitioners, but I can give you some news/reports from websites which show how Statistics will have a great role to play as the "Era of Big Data" evolves. I believe these articles will at least give you confidence that statistics should be a good choice.

  1. For Today’s Graduate, Just One Word: Statistics
  2. How Statistical Science Can Advance Big Data Research Projects?
  3. Discovery with Data: Leveraging Statistics with Computer Science to Transform Science and Society
  4. We Are Data Science
  5. [The Era of Big Data] Must-read articles about Big Data

The last one is from my blog, in which I collected some important articles from media and famous webs, like NYTimes, Forbes, McKinsey, Harvard Business Review, etc. You can find some that outline the future of data science field, and the skills one needs for that field. For example, here is the Quote from NYTimes, the words from Hal Varian.

“I keep saying that the sexy job in the next 10 years will be statisticians,” said Hal Varian, chief economist at Google. “And I’m not kidding.”

What most of the articles elaborate is that as a discipline that studies data -"the science of data", the Statistics field is booming at this historical point. So if there is a consensus, these articles would be the signs of it.

Last, as it might appear to you that I am convincing you to obtain a graduate degree in Statistics or Biostatistics, I don't have that intention, though they are great choices as I indicated previously. Any degrees that fit your interests (like machine learning in Computer Sciences) are good to consider, as long as you know you're preparing your analytical and computation skills. You can even learn those skills by yourself through Open courses on Coursera.

  • 4
    $\begingroup$ (+1) Graduates in Physics or Engineering certainly shouldn't be discouraged from going into Data Science, but I'm not sure I'd count reading those subjects among the most direct routes, for someone who's setting out to go into it. $\endgroup$ – Scortchi Nov 12 '14 at 15:11
  • 1
    $\begingroup$ +1 Notice that the statement "The main issue, if there are, that impede Statistics to promote data science is that people from Statistics are not well trained for large-scale computation and efficient programming" makes the implicit assumption that data science = big data. I would argue that many (most?) problems are either incapable of generating data by the billion or the increased sample size -which all too often is confused for the population- contributes little to solving a problem and may actually hamper the analyst ability to detect a signal. $\endgroup$ – Thomas Speidel Nov 12 '14 at 20:52
  • $\begingroup$ Good points, @ThomasSpeidel. And in fact, a lot of scientific or industrial problems are of small sample size, like experiments in pharmaceutical companies, and many others that need people as subjects. In those situations, definitely "big data" techniques are less applicable. The rise of Big Data is really the result of the revolution of mobile internet and social media on internet. With that said, most of Big Data applications are related IT area. That's why Computer Sciences people now dominate the so-called big data era. $\endgroup$ – Aaron Zeng Nov 12 '14 at 22:15

If you get an MS in applied statistics (followed perhaps by PhD) and get a very strong computing background you won't go wrong. An MS or PhD in Biostatistics leads to a great job pipeline, and if you end up not liking biomedical or pharmaceutical research you would still qualify for a non-medical applied statistics-related field.

  • $\begingroup$ Thanks for this advice. Since your post I've been looking around more at biostatistics departments and one that's particularly appealing is the Biostatistics program through Indiana University at IUPUI (since it can be completed part-time and I could gain work experience while attending). Do many statistics/biostatistics Ph.D. programs offer the option to attend part-time or is this a relative rarity? $\endgroup$ – user60305 Nov 25 '14 at 0:04
  • 1
    $\begingroup$ From what I know it is rare. $\endgroup$ – Frank Harrell Nov 25 '14 at 3:48

The chilling reality is, statisticians are not recognized as data scientists. So while getting a degree in statistics will definitely equip you well for data science, you may not get as many opportunities as you'd think the name of the major would imply.

I don't have the canned answers for you (and nobody does... except Hal Varian, and you may want to talk to him directly -- if you can't Google his contact info, you should not be considering a career in data science :) ). My two cents for you to consider would be:

  1. A program in computer science, with a minor in statistics. A computing degree per se will not equip you well in data science, in my opinion, as what statisticians see in the "data-science-without-statistics" is that data scientists end up reinventing statistics. Hence you will be better off learning it properly to begin with.
  2. A Professional Science Master program in analytics (Rutgers, NC State -- not that I endorse these, just give you examples). The Professional Science Master's programs combine about 60% of the credit hours from science curriculum with about 40% from the business curriculum. I wish I had an option of taking this degree when I was in grad school. Of course this assumes that you can afford it -- you can go to most Ph.D. programs and get full financial support, but you will have to pay for a Master's degree yourself.

Browse Academia.SE for more pointers as to how you can structure your post-graduate training, and what kind of degree you might or might not want. Interestingly, Data Science.SE is currently (Nov 2014) in a beta status, and it is hardly a happy healthy beta.

  • 1
    $\begingroup$ Why no mention of economics, @stask? $\endgroup$ – Dimitriy V. Masterov Nov 12 '14 at 5:09
  • $\begingroup$ The OP asked about [some sort of] science. Economics isn't a science... although it definitely is a well-paid occupation :) $\endgroup$ – StasK Nov 14 '14 at 1:39
  • $\begingroup$ It strikes me as overly harsh to say that "economics isn't a science". I certainly agree that some parts of economics are less scientific than others, though. $\endgroup$ – gung Nov 15 '14 at 17:21
  • 1
    $\begingroup$ +1 for both of your points. A lot of my co-workers went through the NC State program. I'd add that a computer science M.S. in a program that's machine-learning-heavy would also be a good option -- assuming the OP is reasonably skilled at programming, which they'll need anyhow to be a "Data Scientist". I think the bias against statisticians is the stereotype that they are more theoretical or that they use Stata, etc, to work on smaller-scale problems. In terms of economics, I have to say that my favorite stats bloggers are economists. $\endgroup$ – Wayne Nov 15 '14 at 17:40
  • $\begingroup$ Wayne, it is a funny thing to say that statisticians use Stata. R is better suited to work with data of weird structures, and SAS can read and write huge datasets without any modifications to syntax, but Stata is no worse a tool than these two. I am biased, though, although I thought it was obvious in my posts. gung, in terms of Kuhn's structure of scientific revolutions, given the amount of debate in economics, it has not quite formed itself as a science. $\endgroup$ – StasK Nov 16 '14 at 16:11

If you already have a BS in mathematics and you want to analyze data in the field, then an MS in statistics will do much more than a second math degree. Only statistics, biostat, and OR (not so sure) teach the underlying statistical assumptions for statistics problems. Statistics already teaches more math than you need for analyzing data, e.g., measure theory and large sample theory.

Also, Statistical Machine Learning is firmly in the statistics field. These are the tools in ML that we use to analyze data. The other tools are for managing data.


Opt for PG in data science. This is available mostly in US universities and in Europe. And I know they get really expensive. So, why don't you go for online Post Graduation Degree in Data Science and Engineering which will include Machine Learning. I have one in mind : Great Learning-PG in Data Science and Engineering


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy