Chances for MS programs in Statistics/Data Science/CS

Graduate School
1

Disclaimer: I posted this in another discussion forum and copy-pasted it here. Not sure if its against the rules.

I am senior at a US News top 30 school right now, and the majors that I’m considering in graduate school are Data Science/ Statistics/ Computer Science, which are all related.

I will be applying to the CS programs in Chicago, Columbia, USC, UCSD, UCSB, UCD, UCI, UCR, and Statistics programs in Yale, Duke, Cornell, UCLA, JHU, and UFlorida.

My stats are:

Major: B.S. in Mathematics (Pure) + Specialization in Computing
CGPA: 3.49, Major GPA: 3.55, Math requirement GPA: 3.71
GRE (General): 163 (VR), 170 (QR), 5.0 (AW)
Research Experience: 2010/06 - 2011/06, 2014/06-2014/09 at a research institute. The former was done as part of my high school requirement.
Research Accomplishment: first author in two journal papers and a conference paper in Material Physics.
Note: Research involved a fair bit of programming in C++.

For ECs, I have

  1. Table Tennis: I’m a member of the Varsity team and we qualified for nationals last season. Been in it for my entire college career.
  2. Circle K international: Been in it during my sophomore and junior year.
  3. A business related club. Joined in Sept 2015.

Acads wise, I have taken 17 upper division Mathematics courses, including a Stats/CS/Data Science related seminar that required us to do a final project. Other classes I took include Calculus, Linear Algebra (Honors), Real Analysis (Honors), Complex Analysis, Numerical Analysis, ODE, PDE, Probability, Hyperbolic Geometry, Differential Geometry, Topology, Optimization.
I plan to take Abstract Algebra, Stochastic Process, a CS class on Algorithms, and Mathematical Statistics next quarter.

More about me:
I had a freshman GPA of ~3.2 because I knew most of the content being taught but weren’t exactly well-versed in them and wasn’t motivated to study… I know my GPA is pretty bad, considering that the people around me have ~3.7 overall… Also, I haven’t done any Stats and CS courses yet and that hurts my chances of admission too. On the plus side, I have a Specialization in Computing (which hopefully makes up abit for my lack of relevant experience in CS), and will be taking CS/Stats classes in my remaining two quarters.

Which of the programs do I really have a shot in? I personally am lacking confidence that I can get into any of them. Also, what are some other programs (in CS/ Stats/ Data Sci) that perhaps I should consider?

2

Your GPA is on the low side for many of those programs and ECs don’t make any difference for graduate applications. My guess is that you have a reasonable shot at one or more of the UC schools.

3

Thanks for your reply. Yea my GPA is killing me, and GRE doesnt really make a difference. My only hope was that my research publications and experience will make up for my GPA…

4

Your GPA’s not bad per se, it’s just low-ish for the most competitive programs. Columbia’s CS MS says most successful applicants have a 3.5 or higher. The top schools probably have similar profiles, especially for a newly hot field like CS.

Your biggest problem, though, will be that you don’t have any CS courses yet! In statistics it’s possible to get into some programs with just calculus and linear algebra, but the most competitive programs are going to admit mostly students who have at least taken 2-5 statistics classes as well. For example, Duke’s MSS requires a class in probability, a class in mathematical statistics, a class in regression analysis, and highly recommends a class in computer programming. I think you’ll be fine at most stats programs on coursework but check the requirements to be sure.

But most MS programs in CS are going to want a major or the equivalent in computer science, and even the ones that are designed for career changers or people with more diverse backgrounds still want you to have some CS coursework. I poked around at a few of the programs you’re looking at and taking a few CS classes in your last two quarters doesn’t seem to be enough. And for both types of programs, people who have some demonstrated ability (aka who have taken at least some of these classes before the app deadline, so the program can see how they do) are going to have an advantage.

You’d probably be more competitive in a year, especially if you pull your GPA up a bit and complete those classes.

5

thanks!
Do you know other Stats programs that I should consider?

6

First, allow me to ask… are you planning to stop at a master’s degree and not pursue a doctorate?

I ask this because there tend to be 2 different kind of M.S. programs out there… more traditional academic/research-oriented ones, and more professional development ones… when it comes to Computer Science degrees.

The more traditional M.S. programs in Computer Science tend to be more tailored in preparing someone for later doctoral study and… well… research, with a more theoretical curriculum.

On the other hand, more professional development-oriented M.S. programs in Computer Science tend to teach a more applied rather than theoretical curriculum, and are more intended to be a terminal degree.

It is sometimes the case that the traditional M.S. programs have more stringent admissions requirements than the professional development-oriented M.S. programs.

Boston University example…
As an example, B.U. offers a traditional one (http://www.bu.edu/cs/graduate/masters/) and a more professional development M.S. in Computer Science (http://www.bu.edu/met/programs/graduate/computer-science/).

In the case of the traditional one, the GRE is a requirement for admission. However, the other one has no standardized test requirement.

Johns Hopkins University example…
J.H.U. also offers similar options for a Masters in Computer Science… the traditional Master of Science in Engineering (M.S.E.) in Computer Science (http://www.cs.jhu.edu/graduate-studies/masters-program), and the professional development-oriented M.S. in Computer Science (http://ep.jhu.edu/graduate-degree-programs/computer-science).

Note that, similar to Boston U., Johns Hopkin’s traditional Computer Science master’s program has more stringent admissions, requiring the GRE, while the profession development-oriented program does not require the GRE.

My point is, while your GPA might make getting into some traditional M.S. programs a little rough, it is more than adequate for getting into most professional development-oriented M.S. programs.

For most employers, they won’t distinguish between an M.S. in Computer Science that was more traditional-focused and an M.S. in Computer Science that was more geared towards professional development coming from the same university. It really only becomes a factor if you are planning to pursue later doctoral study or a highly academia/research-oriented profession.

And even then, just because a particular M.S. degree is more professional focused rather than research/academia focused does not mean that earning such a master’s degree prevents you from pursuing a doctorate afterwards. It is just that the professional development-oriented master’s programs tend to be less suited for preparing you for doctoral study.

7

Hi,

Yea if I’m doing CS then master’s will definitely be my last degree. If I do Stats then idk, depending on whether I could find a satisfactory job. Could you give some examples of professional CS programs that you recommend me looking at?

8

There are a lot of options available…

Master of Software Engineering @ Carnegie Mellon (http://mse.isri.cmu.edu/software-engineering/web4-distance/MSE/index.html)

Master of Computer Science @ University of Illinois at Urbana-Champaign (http://engineering.illinois.edu/online/masters-degrees/mcs.html)

Master of Information and Data Science @ University of California Berkeley (https://datascience.berkeley.edu/)

Professional Master’s in Computer Science & Engineering @ University of Washington (http://www.cs.washington.edu/prospective_students/pmp)

Master of Science in Computer Science @ Georgia Institute of Technology (http://www.omscs.gatech.edu/)

Master of Science in Data Science @ University of Wisconson (http://ecampus.wisconsin.edu/online-masters-degree/master-of-science-data-science.aspx)

Master of Science in Computer Science - Scientists and Engineers @ University of South Carolina (https://gapp.usc.edu/graduate-programs/masters/computer-science/cs-scientists-and-engineers)

Master of Science in Computer Science @ Johns Hopkins (https://ep.jhu.edu/programs-and-courses/programs/computer-science)

Master of Software Engineering @ Penn State (http://www.worldcampus.psu.edu/degrees-and-certificates/software-engineering-masters/overview)

Master of Science in Computer Science @ Boston University (http://www.bu.edu/met/programs/graduate/computer-science)

… just to name a few.

9

You know, Penn has a MCIT program in the Engineering dept that has a core boot camp of 6 courses that put you through the basic CS components of architecture/operative systems, data structures/algorithms, basic and advanced programming sort of what most MS will expect of undergrads at bare minimum. Then you take the rest of classes with the general population in MSE. I didn’t see data science but they have machine learning group. But it seems to be for ppl with even less CS classes than you and the others are told to apply MSE. You seem to be neither fish nor fowl.

I just read a first person article that characterized the Berkeley data science program as ‘disappointingly easy’
https://■■■■■■■■■■■■■■■■/pulse/why-i-left-my-masters-program-charles-pensig-1

10

I’m well aware of Penn’s MCIT program. But my biggest qualm would be whether I even have a chance of admission. I have taken two basic C++ courses that teaches data structures and algorithms and will be taking a class on java next quarter, but those classes are too shallow. Oh btw I read that article about Berkeley’s program a few days ago. I’m not really into online programs anyway.

11

But MCIT allows for zero previous experience, so your ‘shallow classes are not the issue’. I don’t see why you wouldn’t be a great candidate for that program.

12

I don’t have specific statistics programs to recommend (there are hundreds, maybe thousands of them), but I do recommend thinking about going to a public university for one - to minimize costs. Also, a good idea is that if you don’t go to a tippy-top university (and even if you do) is to go to one somewhere you may want to live. For a professional master’s, that can be helpful because you can do freelance work or part-time work in the area and make some connections. It’s not necessarily, of course, but it can help.

Remember that also as a college grad, your residency is based on where you live - so you can conceivably go live in a state for a year, working, and then apply to their MA in stats program. It’s not totally without risk but could potentially save you a lot of money.

13

Hi BrownParent, yea I will definitely apply to Penn’s MCIT. I just worried that my GPA would not make the cut…