Masters of Data Science
This is a repost from my blog at leadernode.com. Head over there to read my weekly thoughts on data science and get learnings from my experience in data engineering.
A few weeks ago I wrote up some thoughts on Data Science, and I opined that a lot of companies don’t really need data scientists right now. My argument is not that data scientists have nothing to contribute, nor is it that no companies need data scientists, but that my personal experience has led me to work with a number of data scientists at several different companies, and ultimately to question whether we had data scientists because we really had a clear idea of how they would contribute a positive ROI. I suggested that there are a lot of reasons companies hire data scientists and that ultimately it often boils down to Data Science being a very hot buzzword these days. It is, along with AI and Machine Learning, one of the hottest, buzziest words (or to be pedantic, pairs of words) currently in vogue.
Another very hot word pair is Big Data, but I’ll deal with that one in another post. For now I am thinking about Data Science, and in particular about the abundance of MS programs that have popped up at universities the past few years. I live in Jersey City, which happens to be home to two universities, one of which is right at the end of my street. I’m pretty in tune with their recruiting strategy, because I walk past the electronic billboard that advertises it every single day. So I can tell you that they most definitely offer a Masters of Science in Data Science, which will allow you to “Help forge the future of Big Data.” This program is only a few years old, and as someone who works in the… dare I name it such?... data industry, I took note of it when they first started advertising, and wondered what they were teaching.
More recently I have noticed that the other university in JC has started offering an MS in Business Analytics and Data Science. I could hardly help noticing, since their advertisements are plastered all over the interiors of the PATH trains (for those not familiar with NYC’s metro transit system— the PATH is a subway connecting various cities over here in Jersey with Manhattan). This one promises to “prepare students for careers in the burgeoning field of data analytics.” Honestly that doesn’t quite sound like data science to me, but the next sentence does mention Big Data so I dunno. If I were a cynic I might say that it sounds like a program in data analysis, but Data Analysis isn’t a buzzword, so…
Meanwhile, just north of us, in the one university in Hoboken, they have the Data Science Master’s Program, and if you go a bit west to Newark you can get a MS in Data Science or a MS in Data Science or, heck, you can go to the Data Science bootcamp they’re offering right here in JC. And don’t even get me started on the programs you can attend if you want to cross the river into NYC.
When I lay it all out on one page what it reminds me of is, well, a feeding frenzy. Universities have glommed onto the fact that Data Science is hot hot hot right now, and everybody’s falling over everybody else to get their program onto the… subway ad. I mean, that part makes a lot of sense. Who rides the subway? People on their way to work. What do people think about on their way to work? How much they hate their job, how much they’re underpaid, how interesting and lucrative it would be to switch careers to Data Science. If I was a university with a brand new MS in Buzzword, I’d advertise it on the subway (and if I was in Jersey City I’d plaster it all over the PATH). I have an MBA, I understand the rudiments of marketing, and that’s just being not dumb.
But…
Are these programs really delivering value? That’s what I’m wondering about. I mean, I don’t really know. I no longer work in graduate admissions (yeah, I used to, check my linkedin). I wonder what these programs’ placement rates are like, and whether people graduating from them are getting work as data scientists, or data analysts, or business analysts, or what. I do however know how often I get emails/linkedin messages from recruiters who have great data scientists they’re trying to place. And the folks they’re trying to place are not recent MS grads, they’re PhDs (usually) with pretty solid experience. Also I know how many MS grads apply every time we have an open position. Not a position for a data scientist, but for a data analyst or a data engineer (even a front end engineer! It just has to have Data in the description and people apply).
This is not a knock against anyone who is looking for a job. I have a ton of sympathy, having been a grad myself, and I know how hard it can be to land something interesting. BTW I have a BFA in Photography and Cinema, and the closest I have ever gotten to working in either of those fields was when I supervised the research department at a stock photo agency. When I graduated I never thought I would one day be immersed in the world of data and claim to have some expertise in databases and event pipelines. Of course, when I graduated the interweb was in its infancy, so I will cut myself some slack there. The point is: a lot of people are graduating from MS programs in Data these days, and from my perspective as someone who sometimes hires people to be data analysts and data engineers (and occasionally data scientists), it looks like a lot of folks are not really landing data science jobs.
One reason for this might be something I talked about in my post about Hiring for Data Science: a data scientist is, actually, a scientist. They do hardcore research and need to have a pretty broad exposure to different models (like, statistical models) and stuff. A PhD actually comes in handy here. Not that I have anything against the Master’s as a terminal degree, since I have one of those myself, but I would pretty definitely not ever call myself a scientist. I think I have insufficient education for that; I’m sure I could learn to be a scientist, but that would take more years of school and all that dissertation research that teaches you... how to do research. It’s not actually an accident that they make you do a dissertation to get a PhD.
So, again not to diss anybody, but is a Master’s in Data Science really meaningful? I ask this not because Science is in the title. By contrast, I would suggest that a Master’s in Computer Science is a really useful thing and that people with that degree are generally really well prepared to do valuable work (as software engineers— in case that isn’t obvious). But folks with a Master’s in CS aren’t going out in the world to be scientists. They just have advanced expertise in the field of computer science. Folks with a Master’s in Data Science are going out into the world to be scientists. And with some exceptions, I think they are underprepared. Again— not a knock on them. But definitely a knock on the programs that have taken their money and issued them degrees.
It’s not that I don’t think you learn anything in a MS in DS program. I am quite sure that you learn a whole lot. I am just pretty far from convinced that what they have taught you is what you actually need to be a successful data scientist. And yeah, I blame the universities that are offering these programs. Like I said above, it looks like a feeding frenzy to me. Data Science is hot right now, and they’re all cashing in.
So my advice to you, if you’re considering one of those programs, is this: if you’re hoping to be a data scientist, don’t go for one of those programs. It’s going to be very, very hard to break into the field with an MS in DS and no relevant professional experience. If you are already working as a data analyst, and you have done some modelling (in a professional environment), then sure, I could see an MS in DS kicking your career up a notch. I don’t know that you’ll land your dream job right after completing your degree, but if you diligently work towards the job you want, that degree will probably help you.
If you aren’t working in Data at all, and you want to be a data scientist, I would suggest just stopping right there, and taking a long, hard think about your career goals. Do you want to be a data scientist because you are passionate about research and data? Do you have aptitudes for computer science and quantitative work, and you really want to focus on those? Cool, get a PhD in applied math, or physics, or CS. That’s what you need if you’re going to launch a real career as a data scientist. On the other hand… do you want to be a data scientist because all the ads say it’s a hot career with growing demand that pays really well? For you I have four words: Machine Learning on AWS. Data Science is being commoditized, and companies are just not going to be needing to hire people who don’t have deep statistical, CS, and domain knowledge to do it. But that is a topic for a future post, and I have said my piece.
