Profile Eval - Statistics PhD and School Recommendations

[spongbob101]

Undergrad Institution: University of Washington

Type of Student: asian male, International (undergrad in US)

Major: Data Science and Statistics

Minor: Computational Finance

GPA: 3.8 (Major is 3.7)
GRE General Test: Q: 166, V: 157, writing 4.5

 

Research: Some research on COVID-19, no publication. Working on honors thesis.

Teaching: TAing an undergrad course

Experience: Statistics related volunteering, data science blogger

Awards: 1st place in a Datathon (~60 participants). A programming certification by Google

Letters of Recommendation: 1 from research supervisor (professor). I will ask thesis advisor (professor) and TA supervisor for the other 2.

Relevant Coursework: All undergraduate level. linear algebra, differential equations, probability, machine learning, statistical learning, discrete & continuous modeling, database

 

Schools Applying (PhD in Statistics):

• University of Washington
• Carnegie Mellon
• UC Irvine
• UC Davis
• UC Berkeley
• UCSB
• UCLA
• Columbia
• NYU
• University of Toronto
• UBC

Any school recommendations can help, thanks!

Edited September 15, 2020 by spongbob101

[Stat Assistant Professor]

I do not see real analysis as a class in your coursework. Have you taken this class? UW is a great school, but as an international student, your math background seems a bit light compared to most international applicants to Statistics PhD programs.

[StatsG0d]

I agree with @Stat Assistant Professor. To be frank, I don't think you have much of a shot at any of those schools. You should probably focus on the programs ranked below the top 30.

[spongbob101]

4 hours ago, Stat Assistant Professor said:

I do not see real analysis as a class in your coursework. Have you taken this class? UW is a great school, but as an international student, your math background seems a bit light compared to most international applicants to Statistics PhD programs.

Thanks for the reply! I agree my math background is a bit light. I have not taken real analysis. Do you think a strong LOR from a math professor would help a little?

[spongbob101]

3 hours ago, StatsG0d said:

I agree with @Stat Assistant Professor. To be frank, I don't think you have much of a shot at any of those schools. You should probably focus on the programs ranked below the top 30.

Thanks for your reply! Just to clarify, is the top 30 ranked by best statistics grad program on US news?

[Stat Assistant Professor]

41 minutes ago, spongbob101 said:

Thanks for the reply! I agree my math background is a bit light. I have not taken real analysis. Do you think a strong LOR from a math professor would help a little?

No, as it stands, I think you would struggle to get into any top 40-ish programs. Even top 50 (e.g. Michigan State-type schools) are a reach for you.

Your chances would improve (a lot) if you got a Masters in Mathematics or Statistics, where you also take a few math classes, e.g. 2 semesters of real analysis, plus maybe one other math class (e.g. proof based linear algebra, numerical analysis, optimization).

Edited September 15, 2020 by Stat Assistant Professor

[spongbob101]

Just now, Stat Assistant Professor said:

No, as it stands, I think you would struggle to get into any top 40-ish programs.

Your chances would improve (a lot) if you got a Masters in Mathematics or Statistics, where you also take a few math classes, e.g. 2 semesters of real analysis, plus maybe one other math class (e.g. proof based linear algebra, numerical analysis, optimization).

Got it. So for masters programs, what range of schools should I target for?

[Stat Assistant Professor]

21 minutes ago, spongbob101 said:

Got it. So for masters programs, what range of schools should I target for?

You can likely get into most Masters programs with your current profile. It would be better to do it at a reputable R1 university rather than a regional Masters only school, just so the PhD adcoms will be familiar with the rigor of the Masters institution. And it is best to ensure the program isn't an "applied statistics" professional Masters degree, but a Masters program where you are required to take two semesters of Casella & Berger mathematical statistics and theory of linear models. 

The "doing a Masters first" route isn't necessarily ideal, but in your case, your math background is insufficient for most PhD programs in Stat (it might be okay for lower ranked PhD programs in Biostat).

[Casorati]

Your math background is too thin to get into statistics PhD programs. At the minimum you need a course a real analysis and preferably two. I think all programs you listed are out of reach now.  It would be best for you to do a thesis-based master's first because (1) you can secure strong letters through working closely with professors. (2) These programs are generally fully funded. (3) You can take grad level math/stat courses. AFAIK, UBC/Waterloo/McGill have a thesis-based master's. I would take intro to real analysis this term and real analysis II covering metric space next term. If you do well in your intro to real analysis, I think you can get into at least one of UBC/Waterloo/McGill's master's. After completing your master's, you can either stay for PhD or apply to top 20 schools in the US.

Edited September 15, 2020 by Casorati

[Stat Assistant Professor]

Just to clarify, domestic applicants may be able to get away with only one semester of analysis beyond the minimum math requirements and then mostly statistics courses. But since the most competitive international applicants typically have like 3 semesters of analysis (including measure theory, complex analysis, etc.) and the OP does not have other advanced math classes like abstract algebra, etc. that could signal mathematical maturity, the OP should take two semesters of analysis. In order for OP to be competitive against other international students, I suggest they take three upper division math classes minimum.

[MLE]

The issue is at UW it is borderline impossible to get into upper-division math courses without being a math major, and graduate math courses are also seemingly off the table for most undergrads, except the ones who take the honors sequence. At minimum, you need to take Math 327 (was this not a requirement for the major?), then Math 424 (analysis II) and Math 425 (metric spaces) in Spring. These are about the only math courses available to you as a stat major, besides 491 (stochastic processes) which you don't have the background for right now. If you really wanted to push yourself, you could try to take the grad measure theory for stats in Spring since you really only need the material in 424/a bit of 425, but I don't recommend it. An alternative would be to take the advanced calculus sequence, 33X, which will rigorous and very hard but it will give you a far better grounding in analysis than the regular courses and will also open doors for other opportunities in the math department. For instance, I think 33X has some special algebra topics session that you could get involved in.

Chatting with the stat undergraduate coordinator (not the major advisor) is a good idea. Talk to her about your plans if you haven't already. She'll be able to give you more personalized advice.

[Stat Assistant Professor]

3 minutes ago, MLE said:

The issue is at UW it is borderline impossible to get into upper-division math courses without being a math major, and graduate math courses are also seemingly off the table for most undergrads, except the ones who take the honors sequence. At minimum, you need to take Math 327 (was this not a requirement for the major?), then Math 424 (analysis II) and Math 425 (metric spaces) in Spring. These are about the only math courses available to you as a stat major, besides 491 (stochastic processes) which you don't have the background for right now. If you really wanted to push yourself, you could try to take the grad measure theory for stats in Spring since you really only need the material in 424/a bit of 425, but I don't recommend it. An alternative would be to take the advanced calculus sequence, 33X, which will rigorous and very hard but it will give you a far better grounding in analysis than the regular courses and will also open doors for other opportunities in the math department. For instance, I think 33X has some special algebra topics session that you could get involved in.

Chatting with the stat undergraduate coordinator (not the major advisor) is a good idea. Talk to her about your plans if you haven't already. She'll be able to give you more personalized advice.

If that is the case, then I would recommend the OP look into Masters programs where they can take two semesters of NON-measure theoretic real analysis in their first year (along with the usual two semesters of Casella & Berger statistics and applied regression/design of experiments). If they have *no* experience with mathematical proofs, they should certainly not be taking measure theory. 

However, looking at the undergrad major for Statistics at UW, it looks like they actually do require a semester of Real Analysis as part of the major? https://stat.uw.edu/academics/undergraduate/major

However, the OP is a "data science and statistics" major so that might be different than the BS in Statistics.

[MLE]

If OP is a stat major, they had to have at least taken Math 300, which is an intro to proofs class. Last I recalled there were two quarters of Real Analysis required for the major (Math 327 and 424), but they did some restructuring in the past year. I vaguely remember an email about letting students swap the second class for more linear algebra, but it looks like OP hasn't taken that either?

22 minutes ago, Stat Assistant Professor said:

However, the OP is a "data science and statistics" major so that might be different than the BS in Statistics

You're right. Thinking about it more, my money is actually on this applied math degree (https://acms.washington.edu/data-sciences-and-statistics), not the Statistics: Data Science major, which would explain why they had taken no real analysis. If that's the case, OP would have some more options, like taking I think the Optimization sequence (407) or Numerical Analysis (464), neither of which require the proofs class. Although, right now OP is out of luck since these classes are full and tend not to reduce in size. I think OPs best bet is probably to take second year of the honors calc sequence, 33X, since it's basically intense analysis, but that won't help with admissions this year.

22 minutes ago, Stat Assistant Professor said:

If they have *no* experience with mathematical proofs, they should certainly not be taking measure theory. 

All measure theory classes are only offered in Spring quarter, which would give OP two quarters to improve their background. It's common for the honors calc students to take (or try to take, there's never a guarantee they will let you in) the various measure theory classes in Spring.

Edited September 16, 2020 by MLE
some details, formatting

[spongbob101]

27 minutes ago, MLE said:

If OP is a stat major, they had to have at least taken Math 300, which is an intro to proofs class. Last I recalled there were two quarters of Real Analysis required for the major (Math 327 and 424), but they did some restructuring in the past year. I vaguely remember an email about letting students swap the second class for more linear algebra, but it looks like OP hasn't taken that either?

You're right. Thinking about it more, my money is actually on this applied math degree (https://acms.washington.edu/data-sciences-and-statistics), not the Statistics: Data Science major, which would explain why they had taken no real analysis. If that's the case, OP would have some more options, like taking I think the Optimization sequence (407) or Numerical Analysis (464), neither of which require the proofs class. Although, right now OP is out of luck since these classes are full and tend not to reduce in size. I think OPs best bet is probably to take second year of the honors calc sequence, 33X, since it's basically intense analysis, but that won't help with admissions this year.

All measure theory classes are only offered in Spring quarter, which would give OP two quarters to improve their background. It's common for the honors calc students to take (or try to take, there's never a guarantee they will let you in) the various measure theory classes in Spring.

Yes, I am in the acms major. I am taking optimization this fall. But still, I took wayyyy too many ML classes and few math classes. I guess I'll change some PhD programs on my list to MS for a better chance?

[StatsG0d]

12 hours ago, spongbob101 said:

Yes, I am in the acms major. I am taking optimization this fall. But still, I took wayyyy too many ML classes and few math classes. I guess I'll change some PhD programs on my list to MS for a better chance?

I think that's a good idea. Lots of programs have an internal review process for applicants who attended their master's. You might want to consider adding some biostatistics programs as well, as the math requirements are typically significantly less and you might have some better luck with transferring.