MS DS Grad School Advice!

#1

disclaimer: i still have to get into nyu

So I’m currently in the process of deciding which MS program to attend. I have acceptances from Stanford’s MS in Stats: Data Science, Harvard’s MS in Health DS, and the other one I’m seriously considering is NYU’s MS in DS (if i get in, and yes, I am considering this against both Harvard and Stanford). I was a Statistics Major/CS and Biology Minor undergraduate.

I am not fully committed to, but very interested in Data Science/Machine Learning work in the biomedical/biology fields (maybe with innovative health tech startups). In an MS program, I really want to WANT to take the classes in the program, but obviously want to maximize my research and career opportunities as well. I’m really doing this MS because I feel like I have only scratched the surface in learning about new, innovative modeling methods. Of course, the majority of people around me push me to Stanford due to the major Bay Area tech scene and the “clout”, which in all honesty is a solid point, but I question from an employer’s perspective how much this actually matters.

I’m sure the Bay Area connections and opportunities are amazing at Stanford, however I have some issues with the program. While overall I enjoyed my stat major, I definitely saw more utility in the classes that were at least half theoretical half applied (intro to DS, intro to ML, etc.). I never have really enjoyed in-depth mathematical proofs and don’t think I will (struggled with undergrad lin alg proofs a lot lmao). Consequently, looking at the four Stanford core classes’ (numerical lin alg, discrete math/algorithms, stochastic methods, probability theory) lecture notes and homeworks, they seem to be almost fully proof-based. While I’m sure this is all extremely useful in the inner workings of ML methods, I personally think I’ll dread these classes and not get as much out of it as I would hope. On the flip side, some of stanford’s electives are very cool (data driven medicine, neural nets for visual recognition), but only 3 are needed to finish the program, which I think may limit my choice.

Rather, NYU has its own center for DS, one of the oldest in the world, and collaborates with faculty from many departments and has 2 tracks under the DS masters I’m very interested in (biology, biomedical sciences). This program allows for about 7 electives, maybe 4 of which would be for the track, where in 3 others I can explore other interests such as NLP. The classes use python, and seem to have a good mix of theory and application of modern machine learning/DS (they have a freakin class called deep learning for biomedical data). Even when I see the core classes I’m excited to take most of them. Due to the large amount of inter-departmental collaboration, I also feel like it will be easier to find research in an area I’m interested in. But of course the con, it’s NYU, not Harvard or Stanford.

Where Harvard fits in, I think it’s a very interesting program but it’s new and maybe a bit too public health focused since its in Chan. Also, some of the DS classes are very basic.

I know this sounds very one-sided, but its because people are greatly pushing me towards Stanford and sometimes Harvard. Don’t get me wrong, I’m still heavily considering both.

I’m just wondering how NYU is viewed in the DS community, despite less general name recognition, as by far there are more classes there that interest me over any of the other programs. And in general, it’d be great to know what you know about these programs and where they have gotten people. My theory is that I really don’t think the caliber of jobs I attain from both will be very different, and of course school matters a lot less after the first job.

FYI, other schools I’m waiting on are Columbia and CMU.

tl;dr: stanford has the reputation and connections but a lot of classes seem too theoretical for my liking, whereas nyu may not have the name recognition but has classes that interest me more, and harvard, uh…we’ll see.