Stats MS/PhD Profile Evaluation F2021

[Statto]

Hi everyone! I just wanted to post my stats and figure how I stack up! 

Student Classification: Domestic White LGBT Male

 

Undergrad: R1 Large state school, US News top 150

Major: Statistics 

Minors: Math, Economics

GPA: 4.00

Relevant Coursework: (All A, no A+ system)

Math Coursework: Calc I-III, Diff Eq, Linear Alg, Mathematical Modeling(has some analysis)

Statistics Coursework: Stat Methods I-III, Stat Theory/Probability I-II, Statistical Computing, (the following are 4000 level electives):  Biostat methods, Sampling Methods, Categorical Data Analysis, Bayesian Analysis

Computer Science Coursework: Intro to Programming, Numerical Calculus, Security Methods for Computing

Economics Coursework: Intermediate Microeconomics, Intermediate Macroeconomics, Econometrics I

Real Analysis is not offered at my school until the graduate level, I don't know if that is common but I worry that my coursework is a bit weak/too general considering the programs I'm interested in.

 

Research Experience: 

2 years of undergraduate research in machine learning, using Bayesian statistics and estimators. (2 first author publications)

2 Semesters of independent school-funded research in IEMS (poster presentation at conference)

1 Semester of research in data science with professor

1 REU in data analytics

 

GRE General Test: 166Q, 160V, 5.0W

 

Awards: Nothing except purely academic, i.e. Summa Cum Laude, Top Dean's list every semester, etc.

 

LOR's: 1 Super strong, 1 Strong, 1 Average (Based on general estimation.) In order:

From relatively well-known IEMS professor I did my two years of research with.

From REU I did over a summer.

From professor I did my one semester of research in data science with.

 

Applying to: Statistics PhD. Also would be willing to apply to Statistics MS if it would better prepare me for PhD coursework. Honestly, I'm a bit unsure of what I want to specialize in as far as research goes. I enjoy machine learning/optimization, econometrics, and statistical computing. I'm not a fan of biostatistics, so that helps me narrow it down.

 

Schools: CMU(Machine Learning + Statistics), UChicago(Econometrics + Statistics), Berkeley, Duke, UNC Chapel Hill, UPenn

 

Thoughts: 

I think that my coursework is my weakest area, and my research is my strongest area. I'm very open to hearing about other schools that would also fit my interests. CMU is definitely my top choice right now, because I have the most experience in this area of research and their program really interests me. The rest are mostly thrown together through light research and recommendations from professors. Any guidance/advice/evaluations would be super appreciated!

[trynagetby]

I think for places like CMU Stats+ ML and UNC ML it'll heavily depend on the quality of your publications. If they're in ICML, NeurIPS, ICLR, JML,Annals etc I think you have a good chance and should even think more about applying to CS grad programs based on your background.  If they're more applied papers in niche journals like computational basket weaving I think you're heavily overshooting for stats programs given lack of mathematics coursework.

Edited October 8, 2020 by trynagetby

[Statto]

5 minutes ago, trynagetby said:

I think for places like CMU Stats+ ML and UNC ML it'll heavily depend on the quality of your publications. If they're in ICML, NeurIPS, ICLR, JML,Annals etc I think you have a good chance and should even think more about applying to CS grad programs based on your background.  If they're more applied papers in niche journals like computational basket weaving I think you're heavily overshooting for stats programs given lack of mathematics coursework.

Thank you for the reply! I think my ultimate goal is to really hone in on statistics, while being proficient enough in ML and CS to use them as add-ons if that makes sense. That's mostly what my research has been; using ML as a tool to illustrate statistics. Would you (or anyone else reading) recommend going into an MS first so I can get some more math coursework under my belt before going through a PhD application cycle? Or should I apply to PhD's and MS's in the same cycle and see what happens? 

[trynagetby]

If you have the money, I think that having a masters from a top school with A's in 2 semester of real analysis would make you competitive at anywhere minus schools like Stanford. Currently, it seems likea crapshoot but I think its worth applying to PhDs.

Edited October 8, 2020 by trynagetby

[bayessays]

Are you sure there is no "advanced calculus" class at your school? It's hard to imagine a top 150 school that does not offer a class where people learn to prove basic things about calculus. 

[Statto]

1 hour ago, bayessays said:

Are you sure there is no "advanced calculus" class at your school? It's hard to imagine a top 150 school that does not offer a class where people learn to prove basic things about calculus. 

Advanced calculus is a class at my school, though I didn't realize the importance of upper level math classes until I started researching PhD programs (and the profiles of successful applicants). I've mostly just tried to take statistics classes that interest me and seem at least somewhat rigorous, and the advanced calculus route would have caused me to miss some other classes I was genuinely interested in. In spite of that, if I lack the prerequisite experience to perform mathematics at a high level per university standard (i.e. Advanced Calculus or Real Analysis), it may be better for me to look into advancing that knowledge first before applying to these competitive PhD programs. I'm still trying to figure it all out, though.

[bayessays]

 I think you could get into some of the big state schools ranked between 20-50, which are great programs.  I definitely don't think it's worth doing a master's degree and spending all that money.  But I think it's probably unlikely to get into most the programs you've listed (especially Chicago/Berkeley/Penn) because of your combination of lacking advanced math and coming from a non-top school.

[Statto]

8 minutes ago, bayessays said:

 I think you could get into some of the big state schools ranked between 20-50, which are great programs.  I definitely don't think it's worth doing a master's degree and spending all that money.  But I think it's probably unlikely to get into most the programs you've listed (especially Chicago/Berkeley/Penn) because of your combination of lacking advanced math and coming from a non-top school.

Thank you for the suggestions! I'll keep doing my research and figure out which programs fit best for me.