Jump to content


  • Content Count

  • Joined

  • Last visited

About Geococcyx

  • Rank
    Espresso Shot

Profile Information

  • Location
    Research Triangle
  • Application Season
    2019 Fall
  • Program
    Stat & Biostat Ph.D.'s

Recent Profile Visitors

2,297 profile views
  1. Well, let's start with some basics. I'm only graduating undergrad this year, so quite a few folks on this forum have a much more substantial background -- if they comment, I'd listen to them first. This is, after all, a grad school forum, so I'd be remiss in not asking: do you want to go to grad school? If so, do you have any sort of ideas as to whether you'd want a master's or PhD? If PhD, do you have any vague idea as to what you'd be interested in? In statistics, as you may see elsewhere on this forum, research ideas going into PhD's are pretty vague at best for like 95% of us, so it's certainly not disqualifying if you don't have any real idea as to this. Moving past that, statistics and computer science are less and less distinct these days -- I may not like the term "data science", but what that field is turning out to be seems to be a combination of statistics methodology along with topics like algorithm design and database management from CS, numerical analysis from applied math, and compressed sensing from EE (to give a really blanket generalization). As such, I'd suggest taking classes from both statistics and computer science. If you have interests in software design, web design, and such, then I'd point you preferentially towards CS, but a combination of both could be very potent on the hiring market. This advice past the obligatory grad school paragraph, though, assumes to an extent that you want to go right into a job after undergrad. As such, it's important to note that in statistics, it may not be as easy to get a job right out of undergrad as in computer science. Some folks can go right into jobs out of undergrad (see Namita Nandakumar with the Philadelphia Eagles, for example) -- however, many people tend to need master's degrees first, and not everyone has the benefit of going to an Ivy League school like Ms. Nandakumar (UPenn), so that example may not provide a reasonable impression of what you can do. I get the sense that computer science folks are much more accepting of undergraduates entering directly into jobs, although given the massive overproduction of CS PhD's, I'm not clear how true that still is or will be in a few years. Even in this case, though, you are only taking computer science secondary to your degree, so unless you were particularly excellent or prolific in your coding, I'm not sure how much you might benefit from CS's earlier hiring tendencies.
  2. Disclaimer: I'm still in undergrad, this is just what I've seen on the forums. They shouldn't care that much about a lack of bio experience. Frankly, I thought I had at least an OK bio background, and it didn't seem to do me much good. Focus on your math background, you'll be OK! I think genetics is the one area you'd want some bio background, but you can just do other applications of CI instead.
  3. You don't appear to have mentioned what your grades were in Real Analysis -- do you mind mentioning them, along with grades in any other proofwriting-based classes? Assuming that your grades in Real Analysis were uniformly B or above, it's definitely realistic to apply to PhD's next year (certainly so if you got all A's or 2 A's and a B+ or something). That's particularly the case if you mean a public school in the US News top-50 (as I'm assuming), rather than one of the top 50 public schools listed on US News, and also (as I'm assuming) that you're a domestic student. Go check my profile over in this thread: https://forum.thegradcafe.com/topic/117853-2019-applicant-profiles-and-admission-results-for-statisticsbiostatistics/. There are some school prestige and research experience differences that would probably balance each other, particularly with the benefit of actually working in the field for a little bit. I know with my profile I was advised against applying to master's because they'd largely be a waste of my time/money, so I imagine that might hold for you too.
  4. Geococcyx

    Textbook suggestion

    I'm not sure about the later chapters, but the first 6 or 7 chapters appear similar in content to Devore and Berk's Modern Mathematical Statistics with Applications. I'm not quite clear about what sort of difference in content depth there is between the two, but Devore and Berk is pretty friendly in my experience.
  5. Geococcyx

    Fall 2019 Statistics Applicant Thread

    Hey, just wanted to let folks on the waitlists know that I declined my other two offers (out of the three top North Carolina schools), so if you're still waiting on any of them make sure to keep an eye out.
  6. I understand that people probably don't have strong opinions on the topic, but figured I might as well check: who do people think is the best machine learning researcher at Duke/UNC/NC State? They all have some excellent folks who list machine learning as a research area (e.g. Dunson/Mukherjee, Kosorok, and Laber), but plenty of them do a lot of research outside of machine learning too, and it's hard for me to separate out all of their research areas and decide how strong each subject is. As something of an aside, or perhaps a path to alternate answers if people just don't have any opinions on the machine learning research at these schools that they're willing to share (publicly or via DM) -- I usually look for publications in the Journal of Machine Learning Research and IEEE Transactions as a sign of good machine learning research, but are there other journals I should be considering too? Regardless, have a great weekend everyone!
  7. I recall that cyberwulf has usually said it's about the tenth best biostat program (I think that's roughly Columbia/MD Anderson/UCLA territory in US News for whatever that's worth). They have an alum who is the stat/analytics director for the NFL right now, so I wouldn't worry about industry too much. They've had some decent academic placements I think-- I thought they had someone become a professor at Minnesota biostat, although I might be misremembering which department at Minnesota they are in. Given the size of the program, there's not really a big pool of students to be achieving gaudy placements or titles with, so I wouldn't be too worried myself if their top placements aren't quite as amazing as somewhere else that's bigger and has more bites at the apple.
  8. Geococcyx

    PhD: Biostats@UNC vs. Stats@UC Davis

    @Aweaston You probably already saw it, but just in case, there was some discussion of adviser choice at UNC biostat in this thread: EDIT: Oh yeah, statistical genetics is a promising field, and UNC is good at it. I think that is Danyu Lin's primary research area, and he's a Spiegelman winner if I recall correctly.
  9. Geococcyx

    PhD: Biostats@UNC vs. Stats@UC Davis

    Disclaimer: I don't know anything more than you, let alone the other responders. I'd choose UNC biostat myself, but a lot of my research interests are in biological applications anyways. What are your reasons for wanting to study statistics instead of biostatistics? That would help inform our advice. I suspect you could go to industry or academia from either place. If you were looking at biostat faculty jobs the UNC would probably be better, but I don't know how it would play out for stat faculty jobs. If you wanted to work in the FDA, then UNC would be a better choice too. You mentioned a couple prestigious faculty at UC Davis last time you posted who you seemed interested in -- do you have particular faculty you are interested in at UNC?
  10. Thank you! Just to confirm, were any of these opportunities involving collaboration just within the statistics department, or did they all involve outside profs/departments? It sounds like the latter.
  11. I get the sense that part of the experience of being in a biostatistics department is the opportunity for lots of collaborative research -- collaborating with various biomedical research groups at your university, groups in other parts of a school of public health, and collaborating with other professors in your department on both those projects and on projects native to the biostatistics department. Given the nature of large medical grants from the NIH/NHLBI, I understand that such collaborating is pretty much necessary, although it may be a cultural preference in those departments too. I am curious, then, at the extent of collaborative research that occurs in the standard statistics department. Do statistics PhD students work much with professors who aren't their PhD advisers, excepting in statistical consulting centres/cores? If students are working with non-adviser professors outside of specific consulting relationships, are those professors more likely to be in other departments (e.g. biology, hydrology, atmospheric sciences), or do PhD students also work on projects/publish papers with statistics professors who aren't their adviser? I understand this is liable to vary a lot department-to-department, and given that I think most of the current and former graduate students who frequent this forum are/were in biostatistics departments, I'm dubious that I'll hear any anecdotes one way or another about this. Even so, let me know if you do have any knowledge one way or another about this, and have a great weekend everyone!
  12. Undergrad Institution: Big State School, US News top 80 or so overall Major(s): Statistics, Physics, Psychology (applied as a Math major too, decided not to take a non-statistics-related math class) Minor(s): Math GPA: 3.80 Type of Student: Domestic white male, very bland diversity-wise GRE General Test: Q: 170 (96%) V: 167 (98%) W: 5.0 (98%) Programs Applying: Statistics/Biostatistics PhD programs Research Experience: Mostly some social science/psychology research, a little bit of stuff with epi/biostat this year. Awards/Honors/Recognitions: Phi Beta Kappa Pertinent Activities or Jobs: Some public speaking/science communication stuff Letters of Recommendation: Psychology research mentor (department head, did senior thesis with them), genetics professor from a class taught entirely off of journal article discussions, and a professor I had for 1 applied statistics class. Math/Statistics Grades: Calc I (AP Credit), Calc II (B+), Calc III (A-), Diff Eq (A), Linear Algebra (A), Proofs & Logic (A), Intro to Sequences [there were proofs in this, to be clear] (A-), Probability (A), Stat Inference (A*), Real Analysis (B*), a bunch of other applied stat classes that I got all A's in. * indicates that I was taking that class during the applications, hence only schools with late application deadlines (UNC biostat and Columbia stat) or that specifically requested transcripts (Wisconsin stat) got those grades. Any Miscellaneous Points that Might Help: Did SIBS at NC State, profs at my university might have had a few connections they used on my behalf but I don't really know. Any Miscellaneous Points that Might Hurt: Zero grad classes, lots of withdrawals, pretty lacking proofs background, strange choice of letter-writers (no professors from proofs-based classes), really weird personal statement. Applying to Where: Carnegie Mellon - Stat PhD / Rejected (2/18) Duke - Stat PhD / Admitted (2/7) / I might go here, not saying Columbia - Stat PhD / Rejected (3/14) NC State - Stat PhD / Waitlisted (1/30) / Admitted (4/2) / I might go here, not saying Wisconsin - Stat PhD / Admitted (2/22) / Declined UCLA - Stat PhD / Admitted (2/09) / Declined Illinois (UIUC) - Stat PhD / Admitted (1/08) / Declined Ohio State - Stat PhD / Admitted (2/1) / Declined Harvard - Biostat PhD / Rejected (2/15) Johns Hopkins - Biostat PhD / Rejected (2/26) UNC - Biostat PhD / Admitted (1/22) / I might go here, not saying Minnesota - Biostat PhD / Waitlisted (2/27) / Withdrew in early April Brown - Biostat PhD / They said I was on a shortlist and asked how likely I would be to accept an offer, I'm interpreting that as a waitlist (2/27) / Withdrew in early March Pittsburgh - Biostat PhD / Waitlisted (4/2) / Withdrew in early April Based on my profile and my lackluster LoR choices, this seemed like a really top-heavy list of applications to send out (I had some safety school applications sitting on SOPHAS in case I hadn't heard anything by mid/late January). As such, I wonder whether lots of schools just liked me as a quirky applicant, in terms of majors, classes, and maybe personal statement (I know at least one professor did read it, they brought it up apropos of nothing at a visit). Given that I didn't take the Math GRE subject test, in retrospect I might apply to Washington biostat instead of Columbia stat, and I also might replace CMU or Minnesota with UPenn Perelman since I like neuroimaging and CMU just seems really dang tough to get admitted to. As you can imagine, I'm very happy with the results of my application cycle, and given that many of my best offers are from schools that I didn't expect would like me, it's important to remember than your impression of a good fit/likely admission is definitely not the same as the schools' impressions thereof.
  13. I'm just another applicant, and I'm unsure enough with this evaluation that I was pretty hesitant to post, but I can give it a (potentially inaccurate) shot. Your GRE scores and upper-Ivy League background will clearly help you a lot, and it probably doesn't hurt to have a good computer science background in both theory and practice. A 3.5 GPA isn't great, but I'm not sure how much it hurts you at an Ivy -- there's been some grade inflation issues at top private schools, but you're also competing against top students. If I'm comparing against myself, I don't actually know that your Real Analysis grade even hurts much -- I got a B in Real Analysis at a less prestigious school and had no math professors write letters for me, and still got into UNC biostat two weeks after I applied. I am assuming that you're doing more applied research, which ultimately doesn't mean that much, although it's still probably worth having your adviser for that write a letter of recommendation for you. I suspect the more learned members of the forum will correct my assessment later, but I would guess that you'd be competitive at biostat programs 4-10 (UNC through MD Anderson/UCLA level), with an OK but not great chance at the top 3. My reason for thinking you'd be competitive at 4-10 is that I thought I'd be competitive in that range, and clearly UNC thought so too. Meanwhile, you go to a much better school, should have similar GRE scores, and strong computation and coding experience that could make you attractive to higher-up programs even though you have a lower GPA/grades overall (for instance, Harvard biostat in particular seems to care a lot about numerical analysis and computing). That said, I have a couple questions. One -- what exactly does it mean to get "credit" for a course grade? Is that a C? Did you take this class as a pass/fail, and if so, was that required or did you choose that? If the latter, why? I'm super unfamiliar with taking actual academic classes and getting a non-letter grade for them (as you can tell), so I'm curious as to what an admissions committee would make of "credit" in a linear algebra class. If that's standard, then no problem, but otherwise you should definitely have that advanced linear algebra professor be a letter writer. The reason I make that sound like not just fait accompli is because if you have had a close supervisor with a stat/similar PhD or background at your job, you may want to consider them as a potential letter writer. I have no idea if that's the sort of place you work at, though, just something to keep in mind. For stuff you can do, I would say taking more proofs-based classes and getting A's or A+'s wouldn't hurt. Even so, I'm not sure how much you stand to gain by doing that for a year before applying -- you yourself note that the school in question isn't super highly ranked, and you have an Ivy League real analysis professor recommending you already. You shouldn't need to worry about getting involved in research beyond what you're already doing, nor would you need to (say) take a biology or genetics class -- you can learn that in grad school if you really need it. I have a hard time with this profile, because it seems similar to mine except with a lower GPA from a better school, plus better computation and work experience. That said, I also feel like I really overachieved my qualifications during this admissions cycle, so your results may vary.
  14. Just to piggyback off of Zanelol's suggestions, my girlfriend has a disability that she works with her university on, and there's a lot more that they can be doing to help you with regarding your test anxiety -- different testing environments, proctoring, sometimes(?) allowing emotional support animals in to the test. She says that you often need to jump through some hoops to accomplish this, but you already seem to have a working relationship with your school's disability resource center, and your therapist should help you get the necessary documentation. Furthermore, talk to your professor about this. I have a relative who is a professor, and I know that they have administered tests fully orally to students who have issues with written exams. My girlfriend has also had good experiences working with her professors -- nobody wants students to fail or get low grades for reasons that aren't lack of understanding of the material. Also, I took classes over more than 5 years and several summers -- it isn't rare for grad school applicants to take that route, and if it's monetarily viable for you, you should consider it. Finally (as something of an aside), it looks like you might have used your actual name as your username. Just in case you aren't aware, you cannot really delete posts on this forum, so if you are hoping to preserve privacy about these topics in the future, you may want to consider changing your username on your profile/account settings page. An addendum, since I didn't address your questions: if there are any advisers or professors you really like at BC, they are probably better resources for major choices than we will be at a distance. Lots of people change majors, though, so if that is a consideration for you, then you shouldn't worry about having a meandering path. Music majors can do plenty of jobs, but you will have to explain to employers how your music skills will make you successful in that position (hard work, creativity, effective leadership, autodidaction, etc). As for your chances at music after college, the Fine Arts forum on here might be a good place to start, along with your favorite music profs. Sorry I can't be of more help, unfortunately these are questions that require more knowledge of you personally and non-math/statistics fields in general to answer satisfactorily.
  15. Geococcyx

    Fall 2019 Statistics Applicant Thread

    That's an OK idea, but due to the editing times allowed on here, it may make sense to wait until closer to/after April 15th, to get a better sense of waitlist chances if nothing else. EDIT: It's a better than "OK" idea, I got a lot of benefit from looking at those threads too, but I also recall the editing issue cluttering up those threads in the recent past, hence my recommendation.

Important Information

By using this site, you agree to our Terms of Use and Privacy Policy.