Ask HN: Getting a math degree as a working adult?
I am thinking of enrolling in this online BA in Mathematics program through Southern New Hampshire University: http://www.snhu.edu/mathematics-BA-online.asp
I'm a software engineer already working in the field I want to be in for the next 5-10 years (data logging and analysis).
I want to get a degree in math for a couple reasons: a) For me. I've always deep down wanted a degree but couldn't admit it to myself. I would for whatever reason regret it if I never earned a degree. b) When giving people advice around data I would like to train myself around many of the biases we are all so prone to. Learning though on the job in a business intelligence role seems like a very expensive feedback loop. (Imagine making a mistake in a recommendation that costs the company millions of dollars) c) I would like to eventually move into more machine learning and recognize that for now and the foreseeable future statistics and other math will play a HUGE role in that. For jobs in this realm a masters degree is very much encouraged.
Some background on me: My company and myself all work remotely so the discipline that an online degree requires isn't at all an issue for me. Also I am a self taught software engineer so I'm looking forward to taking whatever I learn through my courses and building on top of it as much as possible as I go.
Just to clarify my question is: I've found a math degree to pursue, what (1st or 2nd hand experience-based) feedback can you provide as advice?
87 comments
[ 3.5 ms ] story [ 129 ms ] threadCan you afford it mentally? How much time and effort can you afford to siphon away from your work to excel at your math studies? You can't go full-bore on both.
While some on HN I'm sure have been working software engineers AND pursued a bachelor's degree AND received it in only 4 years (or less, even), the truth is that it is very, very difficult. There will be a lot of misery, and a lot of time & money spent on something that, for now, is over a decade away.
Also since my current role at work is in the same vein what I learn will directly and immediately apply to work do that help lessen the impact at work. Even so it will be tough. Very much so you're right.
I don't understand what you mean by this at all.
That's why maths/physics PhDs are more valuable in industry than BSc's in the exact detailed topic.
Let me give you some advice:
1. Start slow. Take one class to start with. Get used to the pace of lecture/study and getting your assignments done. It's very hard to work full time and also study. You will need uninterrupted space and time to get your work done.
2. If you're taking online classes, you MUST be sure that your teaching fellow/assistant and professor are willing to communicate via email or instant message. If you can attend office hours DO IT. It helps for a professor to see you as a serious student and not as a faceless email address. GO EARLY in the semester to office hours and make it a regular thing. THIS PAYS DIVIDENDS AT GRADING TIME.
3. Make sure your school is a "real" school. Don't go to one of those for-profit schools. Make sure that your school is regionally accredited by the REAL accrediting agencies, not the fake ones that are set up by the for-profit schools.
4. Connect with another student (or students) and work with them. Just like for startups, it helps to be "in it" with someone else.
Good luck.
P.S. One of the guys I helped get through my program was a father of five kids (and one of the way) and he took two flights each way one day a week for a year to attend class. After 3 years he finished. He also held down a full-time job while he did this. It's do-able. You just have to commit yourself to doing it.
EDIT: more advice.
Thank you again!
goes a long way towards helping you figure out whats legit and what's not.
Generally speaking though, you want a program that's "regionally" accredited. Here's a link:
http://en.wikipedia.org/wiki/Regional_accreditation
Just an FYI: I've heard that some schools that were non-profit and independent have been purchased by for-profit entities mainly for their accreditation. Caveat emptor.
FWIW, most "pure" degrees from liberal-arts or engineering schools are probably OK. If your candidate school has a sports program, it's probably OK. Go visit the university. Meet the administrative officials and check out the student body. See if they have a library. There are lots of "tells" that will alert you to the possibility that your program is more about collecting money than offering a valuable educational experience.
Just do your checking beforehand. You dont want to find out halfway through that your degree isn't worth the time or money. Dont forget that all programs within a school aren't created equally. I spent six months doing due diligence on my program and the school is VERY well known.
http://www.scotthyoung.com/blog/2012/07/04/the-diy-degree/
Now, I'm sure he's a smart guy, but I would be surprised if he retains the material long after he takes the final. I highly doubt you can deeply internalize, say the upper level algorithms course in a few weeks (I've taken it) without doing the (difficult) problem sets, or taking the infamous take-home midterm that often has open problems in disguise.
It's difficult to see this as an attempt at truly learning the material he claims to be learning.
I've been looking at Masters courses online recently, but for 9k and very little added monetary gain, i've decided it might be more rewarding to run a side project/business
If you study for a math degree you will probably be a mathematician.
let that sink in for a while because shockingly people discover that later on.
Thanks!
I received a minor in math, and it's been more practical than my major (information systems). It helps you frame and understand the world differently.
Classes get much tougher once you start doing math with letters instead of numbers (usually linear algebra and up), because things transition from computation to understanding the properties of numbers.
I have a degree in CS, but by 30 (26 now) I want to be enrolled in a physics degree, followed by something in the vicinity of aerospace engineering/robotics.
This is not because I think it will be useful for work, just because they're subjects I'm interested in and I wish I had time to be on the cutting edge of these fields. I aim to support myself + my family through some sort of SaaS product(s), so working hard on that at the moment.
Either way, best of luck!
The fact is, without the degree you cannot seem to get into a job that you like, which is about programming. I have learned unfortunately that this is true within a company I work - and I am sure the prejudices will be a lot more if I apply outside.
To elaborate, I have an advanced degree in statistics, but none in comp science. What I have is though interest, and I know I am better than the lot who typically apply for such jobs. I approached this programming oriented team, and the senior leadership were thoroughly impressed - some people in that team (at senior position than mine) actually agreed that they cannot do what many things that I have already done and showcased.
Then I did an 'evaluation' project with them, again in a month I created something that was beyond what they would expect from their own. In the end when I was told that they cannot hire me now because of some leadership decisions - as the current objective is to procure 'x%' of people with com-science degree in the team and that leaves no room for my background.
Though I doubt if I will approach that particular team again (this move frankly speaks of their guts - and I am utterly doubtful of how they can defend me in the future too if I join), I think hiring PhD's may be the norm in the industry - e.g. in other farms like Google too.
I am happy in what I am doing, but I don't like doors being closed to me just because of the background - and there is a chance that I will be happier in a programming oriented job. So, even though as of now it is a far away thought, is there a way for me to get a PhD while working in my field?
It sounds like you're on track though. Keep at making personal projects and applying for any programmIng job you think you might enjoy eventually you'll find he intersection of the right portfolio and right opportunity. Good luck!
I got two years of college under mt belt and had to leave school for many reasons. Went to another university a year later to try and finish and got caught up in the building Internet boom/bust. The experience I gained was VERY valuable and it was definitely the right decision to leave school to pursue my career, but I had the nagging sense that I was missing out on some jobs because I didn't have that degree.
So I found a company that would help me pay for part of my education via tuition assistance and had relatively regular hours and went to it. When I finished the degree, I changed jobs and got a 60% pay raise.
Mind you, I already had a six-figure income BEFORE the degree.
Since then, I've changed jobs twice and each job has been more rewarding and better paying that the last. I definitely feel like my program (which was from a well-known school) helps to open doors. I feel like the degree added a multiplier to my efforts that helps me to earn credibility and break through barriers.
Like it or not, the lack of that degree will subject you to prejudice you wont be fully aware of until you're freed from it. You can certainly be successful without the degree, but having it makes enough of a difference that I'd be hard pressed to recommend someone go without it unless they were clearly an exceptional individual with a clear goal for what they wanted to do in their lives.
Keep looking.
Now, knowing a few bits of CS might be handy, or fun. But you can self-study. If that's not good enough for an employer, find a better employer.
That said, have you considered a Master's degree instead? I know quite a few people who did those in their spare time but I know of no one who has completed a Ph.D without being involved full time. The Ph.D is more important if you are interested in doing research in computer science, not if you want to be a programmer.
Mostly because next year they are going to outsource it all to a cheaper 3rd would country as soon as some "management-science" consultant claims that it's cheaper
and to be honest, with "outsourcing" i think "an X that can program" is going to become even more valuable than "ordinary" programmers, not less.
it sounds like you've been very unlucky.
(i have put a lot of time and effort into making sure that i know the computer science - and the practical parts - too. but it sounds like you understand that.))
How much math have you studied? How much CS? Do you like proofs?
A degree in math is not just about knowing statistics and algorithms. You need to have the dedication to plug away at a seemingly dense problem, without visible success, for often hours at a time. If you've done a lot of programming, you might have the skill set for this already.
Beyond that, the best advice I can give to any math student is this: whether or not your teacher checks your homework is irrelevant. Do every assignment, and make sure you understand it. If you don't do so, you're probably in for a rude awakening on a test. There might be a level you can rise to, in HS or in college, without doing homework, but you will eventually hit the ceiling.
I assume an online course is twice as hard in this regard, as it's easier to put on the back burner. Start out taking only one class at a time. Get used to the level of work you'll have to put into it. It will probably be more than you're expecting.
Finally, while I can't say anything about the university you've picked out, I will warn you: a number of my math/science professors mocked the idea of getting a Bachelor of Arts in the hard Sciences, because BAs don't go as deeply into the subject matter as a BS. If you're not planning on a career change into engineering, academia or the sciences, this probably doesn't affect you, and I wouldn't worry about it. Just be aware of the bias.
Some places, like Oxford and Cambridge, only give BAs not BSs - so it's not always a good guide unless you know the institute.
I'm not saying this to diminish the value of a BA, because what their curriculums lack in technical equivalence are often compensated for in breadth and general well-roundedness.
Of course this is not hard fact, but it's a generally correct pattern in American universities.
Sounds similar to UC Berkeley. The do offer BSes, but not in math.
Here are some majors that offer only BA, not BS:
Here are some majors that offer only BS, not BA: And here are some that offer both: The reason the latter three offer both is that you can take these from two different colleges. You can get your computer science degree, for instance, from the College of Engineering, in which case it is a BS, or from the College of Letters and Science, in which case it is a BA.There are also schools where the only difference between the BA and BS is in the requirements for electives outside your major. For instance, I recall one school where a BA required a certain number of units in a foreign language. The BA and BS required the same total number of units and the same required courses in the major--the BA simply required that you include in your electives a foreign language.
There are even schools where the requirements for BA and BS are exactly the same. When you fill out the form where you tell them how you want your name spelled on your diploma, you also check a box to say whether you want it to be a BA or BS.
When interviewing a candidate, you really should place no stock whatsoever in BA vs. BS, unless you know that his school offers both and you know the differences between them. There are just too many exceptions to make any kind of generalization about the two that won't give too high of a misclassification risk.
Are you sure? In my case the difference between a BA and a BS was taking two biology courses and one English course, versus taking two of English and one of biology.
The mathematics requirements for my degree in mathematics were no different.
At most institutions I have been involved with in some capacity (I am now a math professor), there has not been any distinction.
But furthermore, the requirements for a degree are often pretty minimal, often something like "3 semesters of calculus, differential equations, and at least 8 courses numbered XXX". There are usually a bunch of relatively easy classes, even among those numbered XXX -- a lot of these are intended for future high school teachers -- but there will also be some very serious ones.
In other words, the requirements for a degree are not a terribly useful yardstick (and when we evaluate grad applications we expect people to have gone beyond the minimum, even at my not-terribly-prestigious university). I recommend talking to a professor if at all possible, face to face, who has some research interests in common with your background and interests, and can suggest a course of study suitable for you in particular.
Good luck! I'm biased, but if you ask me the subject is worth it :)
Thinking back to my own BA degree in (pure) math, it was highly theoretical, and I don't see how I could have applied any of the upper-division course content to real-life work.
The program you've found seems more applied. That could be a plus. Unfortunately, it also seems really "wimpy" - the course list is massively watered-down version of what a good department would offer. (E.g. instead of "Algebraic Topology" they have "Geometry for Teachers.") That's disappointing.
Suggestions: 1) Can you find a stronger program, perhaps in statistics (to match your interests), or perhaps an interdisciplinary program between math, CS and statistics? 2) Can you find something close to home you can attend in-person, part-time, for interaction with fellow students? You'll get much more out of it. 3) Check out the Math BA requirements at places like MIT, Berkeley, Harvard - to get a sense for what courses a really good US math department offers as well as requires.
Can anyone provide institutions apart from the open university which offer this?
1.The course content looks ok. There is always some sort of crap class that the department "thinks" you should take. For you it is Mat 380 Error-correction Codes. Out of the classes you get to pick, you should absolutely consider "QSO 320-Intro to Management Science". Mathematics departments (I have been through about 4 different ones) do a terrible job of application. None of your classes except QSO 320 will have day to day usefulness.
2.If you have not paid for this degree yet I would encourage you to look up programs for the following departments. Management Science, Operations Research, Operations Management, Business Intelligence, Data Science. They usually have just enough theory to carry you but focus on problem solving methods and decision making.
3.If your future goal is to go on to a masters then to some sort of research based work, then a math degree is a great jumping point. When ever I have come into a practical application course, I crush the theory and have an advantage in the ease of absorbing the knowledge. If you are not really sure about research/MS/PHd I strongly advise against this plan. You will be disappointed in how little people understand the how to leverage math skills.
In general, school is where you pay money for the privilege of doing homework and writing tests subject to the human faults of a non-perfect professor. I would advise you to look at the course syllabi for the following Coursera classes. So much good content coming from there. At least you will be more knowledgeable of the breadth in these subjects. - https://www.coursera.org/category/cs-ai - https://www.coursera.org/category/stats - https://www.coursera.org/category/cs-theory - https://www.coursera.org/course/operations
The funny thing about that list is none of them are from the mathematics category. Math is pervasive but struggles to compete without context. Take courses with context, unless you really want to pursue an academic career.
Error correcting codes are fun! Taking a class on coding theory gave me some insight on compression, cryptography, and reliable data storage/transmission.
1. Although there is a lot of math in computer science, a typical math curriculum will take you well outside of what might be applicable to your goals in computer science. This might not be important to you, but it's something you should keep in mind as you think about why you want this degree.
2. I looked at the curriculum you linked, and, with your goals in mind, it actually looks pretty reasonable in terms of rigor and relevance. Based on your goals, regression analysis will be a critical subject. But, and I don't want to paint this in too negative a color, most undergrad curriculums I know go further that this program does.
3. Unless your heart's already set on this program, or there are other considerations besides what you've listed, I'd actually suggest looking at other programs first. If what you want is a math degree, than this might be a good fit. But if what you want is a degree that will be most helpful in statistics/data analysis and machine learning (which you listed as important to you) I'd suggest looking at either an Applied Math program or a CS degree. I don't know your background, but there are Masters programs in CS that don't require an undergrad degree. You would still need the equivalent working knowledge of a BS in CS, though.
A math degree is not the best way to go about this. (I just got one myself.) It will teach you to think clearly about groups, graphs, metric spaces and all that, but it won't teach you to think clearly in real life. That's something you'll need to learn elsewhere; although it might be true that a math degree will help you to learn.
Also, reading http://www.snhu.edu/mathematics-BA-online.asp, my main thought was "do they call THAT a math degree?" At the very least, I would put a sticker "applied" on that. That is a judgment, but not one that judges one item as better than the other. For your ambition in machine learning, it will probably be tons more useful than what I (and from his reply, I guess philh) think about when hearing the term.
For an impression of the difference: With a MS in (abstract) mathematics, you will be able to prove properties of some algorithm or formula from some preconditions, and you may be able to generalize them to some highly abstract terminology (for an example, see http://en.wikipedia.org/wiki/Measure_(mathematics), which grew about mathematicians thinking deeply about what the concept 'area' means) but you need not get any experience in recognizing those preconditions. As an applied mathematician, it is more important to have a good intuition on when an approach works then to have a good understanding on whether/why it works in esoteric cases (infinite dimensions, with less than two inputs, etc)
Apart from the personal satisfaction of being a mathematician you go straight to the top of my interview list. Anyone with the dedication and commitment to do a maths degree on their own is the sort of person I want to talk to
My (admittedly personal and biased) guide is maths/physics and pref. advanced degrees first.
Followed by CS degrees from really top notch schools and then I ask why, if they are so smart they didn't do maths/physics at MIT? Acceptable answer is a complete overarching love of computer science.
Followed by no formal qualification devs - with some evidence that they have read/learned more than simply the SDK they were using.
And last on the list are CS grads from some no-name, anyone can get in, CS school. People who did a CS degree because their school counselor said it would get them a well paid job.
[1] http://news.ycombinator.com/item?id=4275634
I was just highlighting that while many bad jobs insist on a CS degree, many good ones treat a CS degree with some suspicion.
See also http://www.joelonsoftware.com/articles/ThePerilsofJavaSchool...
This is exactly the kind of trap that will send a mathematician off on a few minute or years long tangent.
I know it contributes little; given the context i couldn't resist.
Advice: don't just pursue the math that you think will directly benefit your software career.
I went into my part-time math degree thinking that because I was a self taught programmer the limiting factor in my career was lack of exposure to higher maths.
I ended up discovering that the math I enjoyed the most was the more theoretical stuff in analysis, topology, category theory, abstract algebra, etc. The applied math came easier, and I probably could have self-taught most of it, but the challenging pure math subjects have been most rewarding.
My Story:
I earned a BS in mathematics (and BA in Philosophy/History) as a working adult (late twenties – early thirties, tech support and later developer) with a 1.5 hour commute between work and school.
I didn't have any significant difficulty until Real Analysis. That's usually the "weed-out" course for math majors and for which absolutely none of my programming experience prepared me[1]. I ended up switching to Philosophy/History for a while but luckily got the confidence to try Real Analysis again before completely giving up on the math degree.
I think anyone with an aptitude for programming will be fine in the math courses that are science and engineering-oriented. However, real math is about reading and doing proofs; if you have not done much of that I would try to arrange the schedule so that you end up studying stuff like Real Analysis and Abstract/Modern Algebra exclusively in a semester.
[1] In fact, my programming experience probably hurt me because I was overly focused on the “fine logic” aspect of proof writing rather than trying to understand the big picture.
You and I will have a very similar story. The only difference is that I started working as a full-time web developer at my university after I was already taking classes here... I got the job a couple of weeks after my 30th birthday.
And notice I am an Algebraic Geometer, not a funny CS. http://pfortuny.net
Enjoy the knowledge but do not strive for what you do not need unless you do it just for fun (which does NOT look so to me).
When you say "math degree", I don't think of something that applies to machine learning all that much! :)
Math is a huge field, and undegrad math covers a big swath of it. Have you taken a stats course yet? I would consider just taking a stats course first, and then maybe asking anybody you meet with similar interests in data logging/analysis what the relevant courses are to take.
That said, looking over the summary on that page, it does seem more "applied" than most math degrees, which is probably what you want.
I'm pretty excited by the school. It seems exactly what I want in a BA and seems well accredited as well which can be hard to find in an online program.
Thanks for your input!
But you have to do the work, all the problem sets, and the exam - just having the youtube video open in a window isn't enough
Every semester I would work 9-5, take classes from 6-10, go home, fall asleep, rinse and repeat. I had no time to study, but when I did, I often did other things because I had no energy or motivation. I did this for four years (and I already had 2 years of college prior to this!). It was exhausting.
My only suggestion would be to take a light workload and aim to master the subject. I breezed through most of my classes because I already knew the material. For tougher subjects (e.g. statistics, compilers) I struggled. The material wasn't that hard, but because I didn't take it too seriously at the start (when it was easy) I fell behind due to weak understanding of core concepts. Never fall into that trap. Commit to studying consistently, seriously, and keeping at pace with the class.
I only did this because I wanted a degree. I think it's a piece of trash, but I wanted a little bit of security in life (I was in my early 20s and just getting started). Since you want to learn for learning's sake, commit to that. Don't try and cut corners.
While point a) is a good reason to get a mathematics degree, points b) and c) are not. For point b), machine learning and statistics are much more appropriate than mathematics, and for point c), it is worth knowing that machine learning requires a fairly small subset of the mathematics you'd learn in a math degree. For example, a math degree covers many areas of mathematics (such as a heavy focus on proofs, abstract algebra, complex analysis and topology) that have no bearing on statistics and on practical machine learning. Conversely, a math degree also does not focus on statistics and probability, which are essential for data analysis.
Thus were I in your shoes, I would only study the math that is necessary to understand statistics and machine learning, and would start taking a machine learning course. The only math you need is multivariate calculus, linear algebra, and probability.