bayessays

Yeah, good catch, I didn't notice that that wasn't a dedicated linear algebra course. Was the MS entirely online too? Since you want to do more applied work anyways and thus probably don't want to be a professor, department rank matters much less. Those schools I listed might be on the higher end of where you could conceivably get in. There are plenty of lower-ranked biostatistics programs too - take a look at the US News list. Again, a few more math classes would significantly strengthen your application.

The biggest thing you're missing is a real analysis class. It's not 100% necessary for lower ranked programs, but certainly helps. Make sure to do very well on the GRE Q. I would recommend looking at lower ranked biostatistics programs. They will be more lenient about math courses and you can do more applied research (including things that are related to environmental science or epidemiology/social science). I'd start your search at schools like Iowa, VCU, Medical College of Wisconsin, etc.

You should apply anywhere you want, I'm sure you'll get into some top ten PhD programs.

As long as you can get some good recommendations, I think biostat departments will appreciate the unique background and experience and you'd get into some solid PhD programs as is, especially after adding real analysis and doing well.

What subject do you want to get a PhD in? These are professional degrees so they don't clearly lead to a PhD in any subject. If you plan to get a PhD in statistics, you should just take the required math classes and apply.

If you want to do data science, I would recommend leveraging your current degree into an analytics role and self studying some Python on the side.

I can't offer a lot of advice on master's programs, but if you already have a graduate degree in analytics, why spend all that money on a second master's?

Some people on this forum seem to know people who do things like this, but I'll just give my experience/knowledge with tech companies. If you're doing deep learning or something more CS-related, there are probably many more such opportunities. The ML/Statistics research group at Microsoft has 9 people, and many of them are not statisticians: https://www.microsoft.com/en-us/research/theme/machine-learning-statistics/. Google to my knowledge does not have statistics researchers outside of their normal data scientists (who do things that you might consider research?). Facebook research does seem to have a bigger core data science team, which is more research-oriented. The biggest key to being hired is likely having done the right type of research (deep learning, causal inference, networks, something that they have a need for) and having impressive publications. The compensation is going to blow away anything in academia. Entry level data scientists with PhDs start at these companies at $200k+ total compensation. I imagine researchers make at least that. It is my impression that there are not enough of these positions for it to be a definite career path. There are hundreds of people with PhDs from top 20 statistics programs just working as regular data scientists at these companies.

Another advantage of getting a master's degree before switching programs: you never know what's going to happen that might cause you to not finish PhD at the new program - even if you're sure you will, it's nice to have a degree that makes you much more employable in case something happens.

Hard to tell because there wasn't a lot of overlap in schools I applied to and there was a wide range of program ranks. I dropped out of a ~50 ranked program and was wait listed at a top 5 this year. My academic history is shaky, to say the least, and it really didn't seem to affect me nearly as much as I expected. I wouldn't bet on getting into programs that didn't let you in originally though. If you're in a top 10 program, your profile was probably already really good so is there really a lot more for you to show now? But if you can get supportive positive letters from faculty at your current program, I wouldn't rule it out. I had poor undergraduate grades and doing well in my MS and getting letters from known professors seemed to help a lot.

I have personally done this (dropped out and reapplied later) know people who have done it getting their MS. It might be awkward, but if you've done well and hopefully can get a letter from someone in your current program, it won't be a huge issue. Just be sure to explain it in your statement of purpose. Transferring after the second year is a natural point because you can usually get a master's, and you haven't started research yet.

I have no idea what above user is talking about. You have a strong profile - I don't know anything about UK programs, but you would be in good shape in the US to apply anywhere.

NYU, Columbia, Cornell, and Rutgers are not worth applying to with your grades. I would focus on programs that are not ranked on US News.

You are correct that this is bad advice. Stanford explicitly tells applicants in their FAQ that you should not apply to the MS program as a path to their PhD program.

You could've stopped taking math classes after your sophomore year and still had enough math for top 10 programs. You have enough As now after your junior year, even with the pass/no pass, that programs aren't going to doubt your math ability. Dedicated CS classes aren't a necessity, but being able to do some programming helps - a Google internship will be more than sufficient. I'm not super familiar with CS admissions but have heard some of the same things. I think you can get into a lot of really good stats PhDs that have professors affiliated with CS departments/ability to work across departments. The biggest advantages this way are that you can save a year or two of your time, as well as save $100k by not having to pay tuition for your master's. So the opportunity cost of doing master's could be around $250k, and you can always drop out after two years with MS if you need to, or transfer programs in a worst case scenario.

Absolutely agreed with above. If you think you might even possibly want a PhD. You'll get into tons of top 20, and very very likely top 5/10 programs as is. I don't see a MS helping your profile enough to be worth the time or money.

I think you have plenty of math. The subject test may help for some schools, but I'm not sure whether schools that don't recommend it even look at it. A lot will probably come down to your research letters. Are you first author on these papers and/or significantly contributed and will well-known letter writers write great things about you? It's hard to tell these things. If these are stellar publications and recs, I could see you getting into quite a few places on the list. But admissions are also getting harder and I had a similar record to you in a lot of ways (some shaky undergrad grades, good MS grades, publications) and I only got into 1 of the schools you listed out of 5. I suspect you may be able to do a little better, but I don't think the list of schools you gave has enough safe options for all but the most qualified applicants.

I could see you getting into some of those schools, but I would still consider most of not all of them to be in the reach category. I would add schools at the OSU/UIUC level.

Do you have a good relationship with any of your MAS professors, and did you do any research or projects with them? I'd prioritize getting a letter like that if possible. If you could, for example, get 1 letter from a MAS professor you did any type of project with, a letter from a difficult math or stat theory class, and a letter from a professor you did behavioral science research with, that would be a good combo. I think a lot of people struggle to get 3 strong letters, but I'd prioritize anyone you did research with and anyone you took theoretical classes with.

Luckily, you're still in great shape assuming you have pretty good grades. Some people come in with MS Statistics degrees from top schools, but most people are coming out of undergrad. If you take real analysis, and you've had probability/mathstat in your masters program, you have more than enough to be an attractive applicant to PhD programs. The master's, even if it was mostly applied, still let you take the theory sequence and gave you more insight into why you want a PhD, which will be attractive to programs. I'm not sure how high you are aiming in terms of rankings, but if you want to do applied stats, lower-ranked biostatistics programs would probably be ok with you not having analysis and you could just apply now.

The most-used rankings are the US News rankings: https://www.usnews.com/best-graduate-schools/top-science-schools/statistics-rankings These combine statistics and biostatistics programs, so they're intertwined. I used "top 10" as an arbitrary approximate ranking referring approximately to the programs on that list from Stanford through UPenn (#12) For reinforcement learning/bandit/policy stuff, I'd look at Harvard (Susan Murphy), NCSU (Laber), UNC Biostatistics (Kosorok), UW (Luedtke) off the top of my head. For ML more broadly, CMU has a stat ML lab group. You'll find people doing some type of ML stuff at pretty much any top department, though it may become rarer and less theoretical as you go down the list. I'd take the time to look through faculty at any school that might be possibly interesting to you, as it's the only way to make sure you don't miss anything. MIT, like Princeton, doesn't have a dedicated statistics department but has some people doing stat ML in other departments. Stanford has a lot of people doing LASSO/compressed sensing stuff that was founded there, but stats departments in general have a broad range of research going on, and ML is done in many, so I don't think it's very useful to talk about ML subfield rankings especially since it's hard to draw the line between stats and ML a lot of the time.

You have a strong background from an Ivy - one or two Bs are not going to be a dealbreaker. I probably wouldn't bother with applying to Stanford, but I could see you possibly getting in pretty much anywhere else. Could see you getting into some top 10s, and you'll get into a lot of good programs between 10-50. I'd focus on looking at program faculty and narrowing down your research interests a little more. ML is pretty broad. There are not a ton of people doing deep learning stuff, for instance, but you can find them at some departments. High-dimensional statistical ML (like LASSO type stuff) is very common though, if you consider that machine learning. Some people work on reinforcement learning for clinical trials. UT-Austin has people doing Bayesian nonparametrics. These are all very different things, but are all "ML."

I imagine you'll be in pretty good shape. It would be ideal if your GRE-Q was a little higher, but I don't think it will be a deal-breaker for MS programs.

A not-insignificant number of people master out and transfer to another program. It's a little awkward, but I know tons of people who did this successfully and went to great programs later. You can always keep that option open.

Course requirements differ from program to program. If you come in with a master's degree, you'll often be able to skip the first year Casella/Berger courses and maybe the intro to linear regression. But you'll still have to take qualifying exams if the program has them, so a lot of people end up taking the courses anyways just to make sure they are prepared for the quals, because the courses differ slightly from program to program. If you already have a master's from somewhere else, you might be able to to cut off up to a year of your PhD, but it will vary from program to program. Very rarely can you bypass more, and often it doesn't save you much if any time at all.