Considering switching programs, advice, profile evaluation? (stat/biostat)

[roborobo]

Hi everyone, when I was applying two years ago I read this forum fairly religiously.  I'm happy to announce I got into a few decent stat programs (and waitlisted from a ton).  I have no complaints about my current phd program but I'm considering leaving with an MS and applying to another one closer to home for personal reasons.  I'm a domestic student and my background has been fairly non traditional.

 

Background:

BA in econ from a liberal arts school, terrible overall GPA (3.3). - almost no math courses to speak of.

Went back to a state university and took calculus, linear algebra, analysis, numerical analysis, probability/stats, a grad level reading course on numerical linear algebra, and a grad level CS course in pattern recognition.  In total around 20 math courses getting mostly A/A+ and a grad level CS and math course.  I also worked as a math tutor/grader and a research assistant for the economics department during this time.  Part of my weakness when I was applying two years ago was that I had a ton of math courses but very few in statistics (only math stats).

 

My current stat department is ranked between 8 - 20 depending on which rankings you use (NRC vs us news) and is a well respected applied program.  If everything goes right I should have an MS by the end of school year.  My grades from the first year are mostly core classes: Stat theory I, II [A-] (at the Hogg & Craig / Casella & Berger level), Stat methods I, II [A], stat data mining [A], computational statistics [A], overall GPA for this year was 3.89.  This semester I am taking Stat methods III, biostatistics and either artificial intelligence (graduate level, CS) or algorithms (undergraduate course).  I worked as a TA both semesters and during the summer and I passed the master's level qualifying exam this summer.  

 

Unfortunately, I don't have any additional research as most first years at this program focus on the core and passing the qualifying exam.  I should have a master's thesis done by summer.  I'm hoping this time around with my more substantial statistics background I can get into some of the places I was waitlisted from last time.

 

The programs I'm looking at applying to are UCLA [stats], UC Davis [stats], UC Berkeley [stats], UW [biostats], Penn State [stats], CMU [machine learning].  I'm mainly interested in applied statistics and machine learning.

[roborobo]

A related question, for my recommendations, should I have three professors from my current phd program?  I was considering including recommendations from two professors here (one from a course I've taken and another with the professor I'm working with for my master's thesis) and the previous economics professor who I worked for as a research assistant.

[Shostakovich]

Arent Penn State and CMU sorta on the other side of the continent? Looks like it's not purely for "closer to home" reasons

 

One of the faculty members (cyberwulf or biostat_prof) could probably give you a decent feedback, but you realize that many schools won't really give you course credit for a lot of what you've already taken right?

Edited August 29, 2013 by Shostakovich

[roborobo]

Well, I guess it should really say be closer to family  .  However, chances are I'm going to mostly focus on the west coast schools.

Edited August 29, 2013 by roborobo