Data Science (First Year)

Guys anybody else applied for the new data science major at UMich?

@indianhopeful Hey, I did. I have been looking for people who have. I was admitted last month and I’m changing majors from IT Management to Data Science. Although I’m a transfer and you’re a first year… just wanted to know your reason for Data Science.

@akk987 Hey, I’ve been looking for other people too it seems like quite a nice program, I’d love to hear your thoughts on it so far. For me the path to data science kinda started with number plates, and from there I took a course online that really confirmed my interest in it. What about you?

Are you doing DS through COE or LSA?

COE prereqs first year is very similar across all engineering majors.

Calc, Physics, Engr 100, Engr 101(matlab + C++)

LSA prereqs (housed in Department of Statistics)

Calc, EECS 183 (C++, Python/iOS at the end), First Year Writing, and maybe LSA distribution / Foreign Language

Personally, I wouldn’t do DS major, although its not a terrible choice.
I’d rather do Computer Science and take EECS 445 (Machine Learning), EECS 485 (Web Systems), Data Mining, Natural Language Processing, Distributed Systems, or Self-Driving Cars as electives because some recruiters have a hard HR requirement of “Computer Science only”. Startups and non-top tier companies may be more flexible. You would need at least a master/PhD in Math/Stats/CS to be considered a real data scientist anyways.

https://www.eecs.umich.edu/eecs/pdfs/news/3360_EECS_498_Distributed.pdf
https://www.eecs.umich.edu/eecs/pdfs/news/3390_EECS_498_Cars.pdf

DS is very new. Up until now Math/Stats/CS majors were the best choices because there were no DS programs. As schools like UMich, Yale, MIT, etc. develop high-quality programs - these will be the real data scientists.

@indianhopeful Hi, yep I became interested in it after I took some statistics classes and data algorithm/machine learning instruction and classes with Women Who Code. From there, I thought I’d like to pursue this. And I understand that this program is something that requires probably a Masters of Ph.D. to become a real “data scientist”. But I value the program because it gives you the best of both worlds CS and Statistics. I’m someone who really enjoys the data crunching part of it so it’s the datasets and making sense of lots of information that appeals to me, hence why I don’t want to pursue a computer science- hardcore programming languages intensive degree.

@betoh If I get in I’ll probably be going through the LSA statistics track, but I might pick up another major or something along the way in another college. Yea I’d thought about taking computer science with a few data-related electives, but I think by the time I’m out of college and a masters program there’ll be enough opportunity without taking computer science.

@akk987 Ah that’s pretty cool. I have similar thoughts on the data crunching part, and yea intensive programming and computer science alone doesn’t sound like as much fun.

@akk987

DS is a hardcore programming languages intensive degree. Im not sure why you would think its less when its the opposite.

Its even harder than CS because you need to become good a both at math and programming. Hence why CS-LSA only requires up to Calc 2. And DS-LSA requires Calc 3, Linear Algebra, Stats 413, and a Machine Learning course.

You will need strong CS fundamentals for Machine Learning.

@betoh I think I may have presented my opinion in an odd way. What I really meant to say is that I just came out of pursuing a 4 year IT degree program which had a lot of programming and that’s it. I pursued it for 2.5 years and finally thought to myself that I might be getting good grades in this, but I don’t see myself doing one thing only: programming. Correct me if I am wrong but isn’t data science multi-dimensional? I don’t have to be programming all day long if I don’t want to. I loved math and statistics but stopped because I didn’t need much when I was getting the ITM degree. Now that I have started taking math classes again like Calc 3 and linear algebra and stats, I realized I missed it. Hopefully, I haven’t made the wrong choice of pursuing my interest in data science, but if I did, I’ll survive and find my way in the end.

Programming is not leaving any degree I pursue in IT or engineering or data science. I know that’s for sure.

@akk987

You’ll be doing plenty of programming with SQL, Hadoop, Spark, Python Numpy/Pandas/SciKit, TensorFlow, R, MongoDB, and visualization in D3.js in your career. Thats only some of tools. Conceptually the programming involved is more complicated than the CRUD stuff most devs with a CS degree do (except for maybe systems level programmers)

Being a real Data Scientist is not just doing problem sets all day. You need to build models to simulate an infinite number of scenarios and data, hence why you need programming.

Either way, the undergrad will not be enough to be a real DS. You can still be a glorified business data analyst if you dont want to program forever.

The DS major comprises at least 80% of the CS major.

Realistically DS = CS + Stats minor

Or roughly equivalent

I got deferred in the EA round :confused: Let’s see how it pans out

@betoh

As somebody who will be interning in DS (and current Michigan student), I disagree that the programming involved is more complicated than actual CS dev work. I’m not counting web dev, most mobile app dev, and similar stuff as CS dev work because you don’t need a CS degree for that stuff. You don’t need a CS degree for most CRUD stuff either, and both DS and CS people go for these jobs anyway.

I think comparing the two majors in terms of which one having “more complicated” programming conceptual knowledge is the wrong approach, as both majors can be customized to reach a certain level of skill. However, if you really want to go that route, CS majors make up most of the people taking the very niche (and difficult) “programming” coursework involving Artificial Intelligence (and ML and DL) or Operating Systems, which is objectively more difficult concept-wise than almost any of the work you do with data analytics with Python, R, Hadoop, SQL, etc.