Stats PhD Profile Review + Suggestions for schools to apply to

[dymense]

My goal was to apply to Stats and CS departments and emphasize my strong desire to collaborate interdisciplinary on applying statistics/ML to problems in the physical and life sciences. I really loved that kind of work over the years of undergrad research and even now as I work post-grad on research on biomedical imagery. I graduated in 2020 and right now am in a gap year doing research with my old school.

Student Type: Domestic Asian Male

Undergrad: Top 30 University (USNews National Rankings)

Major: Physical Chemistry and Statistical/Data Science double major

GPA: 3.55/4.00

Relevant Courses: 

I don't know if this is a worthy not, but I also took a slew of upper division and graduate classes in Chemistry, Physics, and Biology, including and not limited to: Statistical Mechanics, Quantum Chemistry, Advanced Organic Chemistry, Mechanics, Molecular and Developmental Biology, Biochemistry, Medicinal Chemistry, and Inorganic/metallochemistry 

Math: Linear Algebra, Multivariable Calculus, Ordinary Differential Equations, Discreet Math

Stats: Probability theory, Regression Models, ANOVA, Time Series Analysis, Statistical Machine Learning, Bayesian Data Analysis, Statistical Experiment Design

CS: Programming principles in Python, Introduction to algorithms in C++, Introduction to data structures in C++, Python for Machine Learning

GRE: In-Progress, haven't gotten my score yet

Research:

• 3 years in a research lab working with the chair of the chemistry department, did polymer/materials science research, two second-author publications in two different journals with impact factor of about 4-4.5. Also helped develop a simulation program in Mathematica.
• Currently doing post-grad research with the MechE department making a tool that assists in the classification of biomedical images (I don't expect a publication soon, Grad student said he'd put my name down maybe 3rd author when the paper on the project is finally published). 
• I am also trying to put my chem lab skills to good "humanitarian" use and took a part time job processing COVID19 tests (A friend reached out and said that the genomics lab he works for needed extra bodies to process samples)
• Selected for and presented as keynote project in school data science fair  

Letters of Recommendation: 

1. Chair of the chemistry department (Extremely Strong) - For 3 years he directly oversaw the research I was doing and saw my writing/data show up in the final manuscripts that were submitted and accepted. 
2. Collaborator (PI) on the chemistry project from a different university (Less strong than #1) - they saw me do the legwork for the project, collecting data and presenting it over 2-3 years.
3. MechE Professor (Strong) - I'm doing work with him postgrad and he said openly in group meeting that he loved the work I was doing and was amazed at the progress I had made in a few short months. He said he would love to write a strong rec.

Statements/Essays:

I'm trying to craft a story of someone who has tried really hard to discover that I finally loved the application of statistics/computing to physical and life sciences. My UGrad career was a wild ride of intense chem, physics, bio, math and stat classes and while trying to discover what I really wanted, I realized that I wanted to research something that mixes all of them.

I wanted to get some suggestions on which programs to apply to with my background (low tier, target, reach). Is my GPA too low to get into top programs? How well would I have to perform on the GRE to be considered at top-tiers? 

Edit: also I see a lot of people took real analysis classes, is this super mandatory to be considered a top applicant?

Edited August 11, 2020 by dymense

[BL250604]

To me the overall GPA is less important than how you did in your math/stat courses. The ones you listed are what adcoms would like to see, with the exception of real analysis but good grades and letters can overcome that. How did you do in the classes you listed? Once we know that, it's easier to make some application recommendations.

Also, don't stress too much about the GRE. Your GRE score won't get you in places, it will just be used to filter you out of the acceptance process. Usually if you score in the low-mid 160's, you've done enough. A perfect score may look nice but it won't bump your profile very far in my experience.

BL

[StatsG0d]

Without real analysis, you might have better luck at some biostats programs than stats programs. Even with real analysis, your math background is very minimal. It might be hard to get into any top-30 stats programs.

[dymense]

10 hours ago, BL250604 said:

To me the overall GPA is less important than how you did in your math/stat courses. The ones you listed are what adcoms would like to see, with the exception of real analysis but good grades and letters can overcome that. How did you do in the classes you listed? Once we know that, it's easier to make some application recommendations.

Also, don't stress too much about the GRE. Your GRE score won't get you in places, it will just be used to filter you out of the acceptance process. Usually if you score in the low-mid 160's, you've done enough. A perfect score may look nice but it won't bump your profile very far in my experience.

BL

These are my grades in those classes! I had some feeling that I should target 20s-30s range, am I completely off base here?

Math: Linear Algebra (B+), Multivariable Calculus (B+), Ordinary Differential Equations (B+), Discreet Math (A+)

Stats: Probability theory (A-), Regression Models (B), ANOVA (A), Time Series Analysis (B), Statistical Machine Learning (A), Bayesian Data Analysis (A), Statistical Experiment Design (A)

[dymense]

3 hours ago, StatsG0d said:

Without real analysis, you might have better luck at some biostats programs than stats programs. Even with real analysis, your math background is very minimal. It might be hard to get into any top-30 stats programs.

Ahhh I didn't mention, I am applying to biostats programs. Does this increase my chances here? Could I apply to top 30s if I do biostats?

[bayessays]

I think with the hard science and stat background you'll be more appealing to biostat programs.  UCLA and MD Anderson biostat are considered 27th on US News and I would probably start your search from there, going down. I don't think top 30 is realistic for the statistics programs.  Taking a hard math class like analysis and doing well would help.  You don't absolutely need analysis to get into a good stats program, but then you need to make up for it with higher grades, relevant research, etc.  I think biostat programs will like what you have though and there are a lot of good programs outside the top couple. Considering your interests, I think you'd be happier in the collaborative biostat environment anyways. 

[dymense]

Thank you to everyone who responded to my post, y'all have been helpful! 

2 hours ago, bayessays said:

 Considering your interests, I think you'd be happier in the collaborative biostat environment anyways. 

I also believe that this is the best for my happiness in the field, interests and future desired career path. I am definitely applying to all biostats and biomedical informatics programs now! It now sounds a little naive but I thought that my lacking real analysis as said would not be a major detraction on my application for the statistics PhD (thanks @StatsG0d). My school list is now more realistic with plenty of schools in the 30s-50s (like UCLA, UIUC and UPenn) and a couple reaches in the 10s-20s (like Stanford BMI and UCSF BMI). Thanks for all the advice everyone