bayessays

You have an extremely typical profile. I think one letter from a boss would be fine, but you should also get a letter from a math professor and ideally your REU/research profs. In addition to that, your biggest concern needs to be improving your GRE Q score and getting it up to a 165.

You really can't go wrong - both places have many great professors and you'll be set up for success either way. Only you can decide whether you think research fit or weather is more important to your happiness the next 5 years. A lot of people change research interests as they go along, and I'd encourage you to keep an open mind because you can always pivot research areas during a post-doc.

Not having undergrad profs is not a big deal - I've had profs, research advisors, even bosses, but they all had PhDs in stats. The bigger concern would be whether your recommenders are qualified to comment on your mathematical ability.

I don't think you necessarily need a master's, no. You have all the math and amazing grades/test scores from a good school. I would definitely apply to PhD programs.

Do you have a professor in mind at JHU? As far as I can see, they have about two options for your advisor, Younes and Naiman. Seems like a big risk. JHU has had some successful statistics students, but the one I'm most familiar with was a student of Minh Tang, who left for NC State. MSU is known as having a lot of probabilists, but they do have a lot of people working on stats stuff so it's not like they only do probability (it's just they ALSO have probability stuff, and you have to take quite a few probability classes). If you want to be a statistician, I'd probably go to MSU.

I don't think there is a big enough reputation difference to make a decision based on that. OSU's fellowship sounds appealing, but I would look at housing costs too. A nice apartment in Ames will be much cheaper than in Columbus. As for the location, that's obviously totally up to you -- Iowa State is a big university so it's not like you're in a really small town and there's plenty of places to eat, coffee shops, things like that. Des Moines is less than an hour away. But if big city life is important, then obviously Columbus is the way to go. Also look into research/prospective professors, of course.

I don't think you can go wrong with the advice to "take more math", but I also don't think you can go wrong with the advice "make sure you don't do poorly in math classes." Taking this class will, in all likelihood, not make or break your PhD admissions but getting a B or C will hurt. Since you already have some analysis background, "doing analysis" as it relates to stats/ML is essentially using the ideas of convergence etc and applying them to probability/algorithms problems - you'll have to learn these things in the context of statistics anyways and you'd be better off just studying them in context. For example, you'll take stats classes and learn things about convergence in probability -- you'd be better off learning about convergence in probability directly than taking a bunch of real analysis classes in "preparation" to do statistics stuff. (of course, there are exceptions and if you want to do extremely theoretical research and go to UChicago or something, you probably want to take more math, but if that were the case, I think you probably wouldn't be asking this question)

I'd also suggest you ask the department this directly -- "do you support/encourage students doing internships over summer?" is a question that any reasonable statistics department would expect from applicants, and if this is important to you, you should go to a department that enthusiastically says "yes."

I think the 30-50 range on the new US News rankings is where I'd probably recommend applying. Your biggest problem is the low math grades which will be hard to completely overcome. I'd apply to lower-ranked programs and biostat programs if nothing works out this time. Get a job as a data analyst for a year or something.

UNC has been weird about funding forever (I rejected their offer years ago because they wouldn't commit to funding you through the entire program), but others on this forum claim that they improved their funding offers recently. I would bet that they are just figuring out the funding still and will get back to you in a later email.

A lot of business schools have professors doing statistics, yeah. I know some pretty solid ML/stats people in business departments. Psych, neuroscience, etc... you'll find people doing ML/stats. Regardless, though, you have to find those people. You absolutely cannot go into a program and expect that you'll be able to somehow get an advisor from another department. I wouldn't go to a department that doesn't have at least a few people in the department itself that would be interesting to work with. Part of this also depends on what you mean by "theoretical research." What you consider theoretical might be more commonly thought of as methodology or even applied for some people.

Your best bets in NC are probably applying to biostat at Duke. Followed by UNC stat/biostats. NCSU and Duke stats will probably be the harder ones. If you're really wanting to stay in NC, I think UNC - Charlotte or Greensboro have lower ranked PhD programs too. SC has UofSC stats and MUSC biostat. Georgia has UGA stats, which I don't hear much about, and Emory biostat. You got Florida covered from last year's schools.

No, not really. If you do a biostatistics PhD at a top 5 school, it may be possible to do some machine learning stuff, but it's certainly not going to be very theoretical. What you should do is look at biostats PhD programs and go through every single professor's web page and see if they are doing anything that interests you. Look at their papers in machine learning journals. You will have to do the same thing for any stats programs you look at anyways, because ML research is still hard to find in most stats departments.

There is absolutely zero reason to think about your grade in this class unless it is going to interfere with you graduating.

Not only is it common, but it's MUCH more common than academia (and not just as a backup).

1. No need to take more math. And the fact that your courses are at a good state school and UPenn will look good and help you a little, yes. 2. Not necessary - your work experience will be valuable in same way as research, in that it will give you context to think about solving data-related problems. For a masters, this is plenty. 3. Not usually a huge factor. 4. Yes, a little. I would apply anywhere that you can afford to apply if you do well in the math classes. I would also consider seeing if you can get an online masters, perhaps paid in part by work. This would be the most economical option, and given your interest in remaining in industry, I think the best thing you could do would be to keep working and also get a MS degree through an online program (Off the top of my head, Berkeley has a MS data science online, and some good statistics programs like PSU, Iowa State, Colorado State and many others have online masters)

If their letters won't be any better/more personal, then I don't think it's worth it. I personally don't think it's a big deal.

I think it'd be worth it. Competition is high these days and I think that's a pretty big signal of math ability, and also will look good in terms of showing you're taking going back to school seriously. Will also help you personally during the program. If you can afford it, I'd highly consider it.

13-15 is on the higher end but absolutely within the range of normal. Most people limit their applications due to financial constraints (since it's ~$100 per school). I'd recommend most people apply to at least 6 or 7 programs unless they have very strict geographic constraints (for instance, you have to stay in one city due to family). 20+ is excessive.

Do not get a second bachelor's, ever. Try to get your GPA up to 3.0 before graduating if possible and apply to low ranked/unranked MS programs that are applied and focus on data analysis.

You can't go wrong. Obviously if you're completely set on FDA and can work with Muller, UCD is a great choice. Better weather too. The personal stuff is very important too, and Michigan is a great program. I'm sure you would be able to find some great research there you would enjoy.

They will probably want an answer pretty quickly because they'll have to offer those spots to other people if you decline. You aren't going to learn anything about the program in the next week that you couldn't learn otherwise, so I think they would expect you to be able to decide within a few days or a week.

UCSD would be a good option if you want to do very theoretical stuff but you won't have many other choices there. Some very good people. Duke biostat would only be a good option for someone who definitely wants to go into industry after. Minnesota is an all around solid department and would be a good choice for someone who is in between those two extremes.

No personal experience, but I think it's been pretty hard to advance recently with an MS in pharma because there are so many PhD graduates and they'll usually just hire those if available. Take this with a grain of salt though if someone with first hand experience chimes in. Data science is a route where you can have great career success with an MS and make more money doing almost the same thing. I'd recommend looking into that.

You just have to look at the individual faculty then - none of those programs you listed have a large core of statisticians actively doing research. JHU had some successful students from one particular advisor. Georgia Tech is an industrial engineering degree and you can't even easily find who is a statistician on their website. Maryland has some theoretical people but most are pretty old, not sure if they are still very active, though they have some younger faculty too. I can't think of a math/applied math department besides Notre Dame (which is on the list) that has enough statisticians to consider attending without an advisor picked beforehand