Thursday, October 6, 2016

Thirsty: Peak CS?

I'm admittedly observing from a biased viewpoint, since most of my teaching load consists of  sections of Writing for Scientists, but it seems to me that more and more of my students are majoring in Computer Science (and, as I noted in a comment on Frankie's post below, they all seem to be obsessed with Big Data, though always with good intentions; whatever majors college students choose, a good many of them seem to have hopes of saving the world, as well as making some money, via the profession of their choice, which is both touching and encouraging.  They also remind each other regularly of the ethical issues involved in collecting data -- also encouraging).  In some ways, this makes sense; computers are, after all, essential to all kinds of activities these days, and my university has a growing engineering school (I suspect a lot of universities have growing engineering schools). 

But it still seems a bit unbalanced, even within the science realm, especially when I consider that 10-15 years ago everybody seemed to be going into nursing and other health professions (partly because of forecasts of a coming shortage of nurses,  which sounded ominously familiar to this possessor of a Ph.D. funded by people who projected a shortage of humanities Ph.D.s, and, from what I've heard, the nursing projections had similar outcomes -- the work was there, all right, but a lot of it got outsourced and/or rearranged so that lower-paid people could do more of it).  It's not as if the baby boomers got younger, or less in need of help/replacement, in the last decade+ (though the CS folks have solutions for that, too: those that aren't obsessed with Big Data are obsessed with self-driving cars -- which of course would be useful to seniors, though students always seem rather surprised when I point that out -- or robots, including care/assistance robots). 

And of course as a humanities proffie (albeit one who teaches almost exclusively gen ed classes), I'm aware of the declining number of humanities majors, even as I see occasional calls for people with the flexible skills built by pursuing such majors, and wonder if the pendulum might just swing our way one of these days (probably long after my retirement). 

But for the moment, I'm wondering whether others are seeing a rise in CS (or similar) majors, and/or see other cyclical patterns in choice of major?  If so, what do you think drives those cycles?  And do they serve students well, or just contribute to the number of un/underemployed, and so dissatisfied, investors in a college education (formerly known as college grads)? 

--Cassandra

20 comments:

  1. My school has removed Physical Education from many of the major's required classes list. Now they are discussing removing Philosophy as well. There is a trend toward only requiring classes for degrees that lead to gainful employment.

    ReplyDelete
    Replies
    1. Yuck. The gen-ed list at my place is a bit short compared to what I had to do, but at least there is still of subtext that an educated person should know at least something from a broad selection of topics across the range of human achievement. Some laboratory science, some human science, some humanities, some art, a PE class and a health class. Some stuff where you have to write and something related to the ways people think and live in foreign places.

      And for all that many of our students—many of whom come in painfully provincial—leave our hallowed hall with a fairly limited understanding of the range of human values and ways of being.

      For FSM's sake let's not let that scope narrow further.

      Delete
  2. The Bureau of Labor Statistics projects an 8% decline in the number of programmer jobs in the next ten years. "Computer programming can be done from anywhere in the world, so companies sometimes hire programmers in countries where wages are lower. This ongoing trend is projected to limit growth for computer programmers in the United States."

    Henry Ford may have been a robber baron and a bigot, but at least he realized that a business needs at least some people out there earning enough to buy the product.

    ReplyDelete
    Replies
    1. Iiiinteresting. Thanks, Frankie!You're always handy with finding the actual facts (or at least projections). I can imagine someone making an argument that our local market, which absorbs the great majority of our graduates, has a disproportionate number of computer/IT jobs that cannot be outsourced, and they'd probably be right, but that still sounds like potential trouble ahead.

      Delete
  3. "Big data" is an outgrowth of large datasets, powerful computers, and extensive computer networks. It's astonishing what's being done in astronomy today. Three automated telescopes with big digital cameras have been measuring how bright half a billion stars are, every 10 nights, for the past 10 years, and all these data are publicly available online. My students and I are having lots of fun picking over them. European Space Agency's GAIA spacecraft is measuring reliable distances for over a billion stars. It’s expected to find tens of thousands of planets of other stars by 2020. A very large automated telescope is expected to become operational in Chile in 2023. It’s expected to produce over 10 million alerts (of asteroids, stellar flares, supernovae, etc.) every freaking night.

    The first time I used a computer to observe the Universe, I got spectra of a star exploding: it was fun. I used an instrument run by a PDP-11 computer, which had 8k of RAM. It was a wee while ago, in 1978.

    As you can see, the subject of “Big Data” is different. (Yes, I know the word “data” is a plural, but I mean the subject here.) With the torrents of data today’s computers can get and work with, humans need computers even just to begin to understand anything. So “Data Science,” informally known as “Big Data,” is a new, burgeoning field, largely about software and statistical techniques for analyzing large datasets.

    There may be jobs in this. Ever shop on Amazon, or see the film Minority Report? Every merchant will want to keep track of your wants and likes, like that.

    So will the government. I know: just freaking great. As I keep saying, I liked computers when they were still mainly indispensable tools for research in physical science, and not much else. That was a wee while ago, too.

    ReplyDelete
    Replies
    1. I think the Big Data bubble may pop soon, not in astronomy or other data-intensive science, but in non-science private sector applications. Ultimately employers will get tired of investing in Big Data if it's only descriptive and not predictive. I know someone who uses Big Data to identify factors that lead to employees quitting their jobs. Companies want this information because the departure of a skilled employee can genuinely affect the bottom line. They want to know who might be poached by a rival firm or leave for other reasons. The software is even supposed to suggest interventions--would the employee stay if offered a raise, more vacation time, or a flexible schedule? I suspect it would be more useful for managers to simply listen to their employees and ask what they need to be successful.

      Students need to learn something in addition to coding. Numbers without human judgment to interpret them are useless.

      Delete
    2. Oh sure, there are jobs in this. But what the students don't get is that they have to UNDERSTAND the statistics (not just push buttons on a mining tool) and they have to understand the bias that might be built into their collected data. It is really cool (and a bit uncanny at times) to see what one can predict, but we really shouldn't be letting machines decide.

      This will be a hype for a while (remember when the world couldn't get enough relational database people? 4GL folks?) The topic burns high, and then there are just a few glowing coals left amongst the ashes.

      Delete
    3. I'm with Suzy on this. The hard part is in understanding both the math and the messy realities that confound the math: the data are about human beings, the metrics are chosen by human beings and in many cases non-trivial parts of the data are reported by human beings.

      That makes is harder than merely tracking decaying particle and their daughters through a fine-grained active detector (which is one of the things I've done with big data).

      Delete
    4. Kids, kids! Anytime old people talk like this about anything new, it evokes the spirit of Lord Kelvin, who really did write and say the following:

      "Trust you will avoid the gigantic mistake of alternating current."

      "Wireless is all very well but I'd rather send a message by a boy on a pony!"

      "It is quite certain that a great mistake has been made—that…geology at the present time is in direct opposition to the principles of Natural Philosophy."

      “The disintegration of the radium atom is wantonly nonsensical.”

      “No balloon and no aeroplane will ever be practically successful.”

      That said, academia does have a long history of overhyping the hot, new thing. How many colleagues do you know who are still making fullerenes, or nano devices? So, sometimes they stick, sometimes they don't.

      Delete
    5. Frod, thanks for the reminder that there is, indeed, some pretty cool stuff that can be done with big data analytics. As is often the case, the tool is morally neutral; the question is what you do with it. There's even some cool large-corpus analytical work being done in my own field, English, but it really needs to be combined with old-fashioned close analysis (and, some would argue, vice versa) to come to anything close to an accurate conclusion. It also really, really matters where the corpus comes from, and what factors may have influenced which texts were preserved, digitized, etc., etc., which gets to the human-level factors Suzy mentions. That's always the one that gets to me, especially when it comes to the ever-popular questionnaire-based data: it matters so much how the questions were worded, who answered them, under what conditions, etc., etc., and often there seems to be little awareness of that. It would be an interesting experiment to set a creative writing class to the task of imagining scenarios around the filling-out of a particular questionnaire; that would probably produce some illuminating results.

      Delete
    6. Yes, Prime Minister summarized the problems with surveys rather well:
      https://www.youtube.com/watch?v=G0ZZJXw4MTA

      Delete
  4. One of my employers was recently given Carnegie Tier One status; its most popular majors seem to be things like business, engineering (or engineering technology for the ones who can't get into engineering proper), marketing, supply chain and logistics, hotel management, and other "practical," non-academic subjects. Meanwhile, liberal arts majors are *required* to get a teaching certificate as a minor.

    ReplyDelete
  5. At my school, the currently trending major is "public health education," which sounds fine until you realize it's basically a dumping ground for anyone who couldn't get into the nursing program. We have an absurd number of public health education majors who are neither scientifically literate enough to really understand health information, nor good enough at communicating to educate people about anything. (One of my colleagues claims that they all become personal trainers, but I doubt there's much of a market for personal trainers in our notoriously impoverished Deep South state.)

    Honestly, though, I don't know what these students could / should be doing instead. The problem is that they are good at neither words nor numbers, and are also not very good with abstract ideas in general, so there ARE no majors that are a good fit, except for the few who are exceptionally talented at something like music or studio art.

    ReplyDelete
    Replies
    1. This is why we need good blue-collar jobs and service jobs. Let's face it, some people are never going to be able to master statistics. Those people still deserve to make decent wages doing meaningful work.

      Delete
    2. Indeed. And the great majority of us like having lights that light, toilets that flush, and HVAC systems that work as expected.

      Delete
    3. "Public health education" sounds a bit scary -- like it's going to generate a bunch of statistically ill-informed people whose recommendations you have to make your way through before talking to an actual doctor. We have a reasonable number of public health majors, but the science/social science requirements seem to be pretty rigorous.

      Delete
    4. I'm not totally convinced that going into a trade is the right answer either, because electricians / plumbers / HVAC repairpeople need to be able to problem-solve and figure out solutions, and for that, they presumably have to understand WHY things work. I think it's a fine option for students who are reasonably bright but not academically inclined, but what do you do with people who get hopelessly muddled when they have to do anything that involves understanding and applying general principles, rather than rote memorization? (Often, these students are not too good at rote memorization either, I suspect because they're treating separate bits of knowledge as discrete entities and can't see how they connect or relate to one another.)

      Delete
    5. Good point, porpentine. The trades do require good critical-thinking (and human-relations) skills, as well as a fair amount of physical strength (and/or the intelligence to figure out workarounds) and manual dexterity, *and* the ability to work independently. That's a pretty demanding set of requirements (and explains why truly skilled tradespeople make good money).

      I'd suggest putting the students you describe to work scrubbing toilets, floors, tables, etc., but even that can be done either badly or well (and there are vulnerable people in many of the places where such work needs doing: hospitals, care facilities, schools, etc. So the ability to see one's job in a larger context is still essential. Many people with very little schooling can do that, and do it well, and they deserve the chance not only to do such jobs, but eventually to become supervisors if they desire and have the appropriate skills. Sadly, I suspect those jobs often go to college grads who do them less well than many of their supervisees could.)

      I'm not big on "follow your passion" as a guide to finding a job; not everyone is lucky enough to find a paying job that is also a passion, or at least a vocation. But for goodness' sake at least find a job doing something you care about, and can do reasonably well, whatever that is.

      Delete
    6. There's research that suggests that if your passion becomes your job, it stands a good chance of becoming one or neither.

      And, passion itself isn't necessarily enough to get paid, even if relevant jobs are plentiful. One also needs talent, coupled with 10 thousand hours of practice in order to attain expertise.

      I'm reminded of a sitcom episode episode in which one of the characters who, having decided to take up dance again, is playing a videotape of her audition for her friends.

      "Ever since I was a girl," she gushed to the aghast audience, "I've wanted to dance so badly!"

      "Looks like you got your wish," said a friend.

      Delete
  6. Big Data is a native part of computer science. Computers were initially built to crunch numbers. They may think they will be changing the world by predicting the next Ebola epidemic or whatever, but in fact they'll be doing what programmers have always done: writing algorithms. And if they are talented, maybe designing algorithms and doing some other cool stuff that doesn't get done by entry-level programmers.

    If they are concerned with employment in the CS field, I'd recommend security over Big Data because it's a concern for every company, whereas Big Data is much more of a niche. Also there are some great opportunities with computer forensics, including the CyberCorps scholarships offered by the government.

    CS is a great field and I recommend it for any student who likes math, but if they want to get up close and personal with Big Data, they should major in Quantitative Analysis or Statistics or plain old math (but math isn't plain! It's elegant and beautiful!)

    The easy way to spot the programmers is to just write a puzzle on the board, nothing too complex, just a simple brain teaser, and see who solves it without being asked or prompted, and then see, at the end of class, if any who haven't solved it have stayed after class to ask you about it.

    The programmers will be compulsive about it because they are addicted to problem-solving. Those who don't care about it can still be good programmers but they won't be great.

    ReplyDelete