As Big Data Plateaus, Data Science Education Grows

gthuang88 writes: Even as the hype around big data has died down, opportunities for data scientists are expanding. Johns Hopkins, NYU, and MIT are among the schools offering courses in data science, and IBM and other big companies are investing heavily in training programs. Now a startup called DataCamp has raised $1 million to expand online courses in R programming, Apache Spark, and other topics. The deal speaks to the opportunity that venture capitalists see in training the next generation of data scientists and business analysts. It also shows online education is specializing beyond platforms like Codecademy, Coursera, and Udacity.

Data ?Science ?

[Crashmarik]

Much more an engineering discipline than a science.

Re:

[Mr D from 63]

I'd call it a scientific instrument. Like a microscope. Scientists learn how to use it, engineers design and build it.

"Data Science"?

[Anonymous Coward]

"Data Science"? Why the fuck isn't it just called "Statistics", like it has been for centuries?

Re:

[Crashmarik]

Well give them something there is the information extraction going on. Don't see how that rises to a science.

Re:

[umafuckit]

This is somewhat fair, but when I think "statistics" I think fitting a GLM to extract some parameters or test for significance. When I think data science, I think of stats but I also think of dimensionality reduction, visualisation, lots of programming, clustering, etc.

Re:

[LaurenCates]

"Statistics" isn't cool enough, and since it doesn't have "science" or "mathematics" in the name, it doesn't fit into the acronym "STEM".

And you can't invoke the mighty need for "STEM" if you have to make people think about how it's related.

Re:

[gweihir]

Because then it would not sound "modern" and "hip". Personally, whenever I hear "Data Science", I think "bullshit", because it really is just statistics, but with big egos and small skills.

Re:

[invalid_user]

Actually, don't worry so much about the ego thing. I teach data science, and I am fairly candid about the "b******t" thing. Many other people I know who does "Big Data" knows and admits it's "b******t", too.

As for the skills part... hmmm, you want to pick a fight with me? :)

Re:

[gweihir]

Well, if you admit that your respective skills are in "bullshit", then there is obviously no need to fight.

Re:

[invalid_user]

No, I do not admit my skills are "bullshit", though the skills required of current "Big Data" stuffs are.

The fight is on? ;)

Combo of CS, Stats, Analytics, & domain knowle

[langelgjm]

While I think this has changed over time, initially I think some statisticians were suspicious of techniques coming out of computer science, e.g. SVMs. And still, machine learning is a rather niche field of statistics that requires a fluency in CS that many statisticians don't have (or need). Check out this discussion. [stackexchange.com]

Of course there are some statisticians who are also good CS people (think Trevor Hastie and Rob Tibshirani). And a lot of stats people have great domain knowledge in their areas. But I think "data science" is supposed to be the combination of stats, CS, practical programming ability (e.g., cleaning and manipulating large datasets, which is definitely not part of traditional CS or stats education), ability to communicate results effectively, maybe throw in some visualization, knowledge of how to query databases, and domain knowledge to interpret what data mean. Also, some types of data (e.g. text with the aim of NLP) are pretty infrequently touched upon in stats education.

That said, I get the sense that a lot of places looking for "data scientists" are actually just looking for business intelligence people.

Re:

[invalid_user]

Yes, and if America can produce enough data scientists, you don't have to H1B-hire SAP "consultants" (shoot me please for saying that) from India.

It's a win(Americans)-lose(SAP(German company))-lose(Indian consultants) situation. :)

Re:

[penguinoid]

"Data Science"? Why the fuck isn't it just called "Statistics", like it has been for centuries?

Because statistics show that you can get more money by calling it data science.

This field has already yielded a lot of nonsense

[Karmashock]

The issue is that if you massage the numbers hard enough they'll say whatever you want. And often what is being done to them is so complicated that it is hard for anyone that isn't very familiar with the specific algorithms to even know what happened.

A big issue here is that it is very hard to audit this sort of science and if you are inclined to sensationalize your paper or make it appear more interesting than it would otherwise... it doesn't appear to be hard to do.

What I hope comes out of this education

Re:

[Mr D from 63]

I agree, and think you articulated it well. There is too much abuse of data, often but not always unintentional. Better education in how to work with data should be a good thing. Even in base curricula, not just science.

Saying about fields with word "science" in title

[blahbooboo]

If a field has the word "science" in the title, it's probably not science.

Re:

[gweihir]

Fortunately then for me that I graduated in "Informatics", not in "Computer Science". Of course, all you people in the English-speaking countries are screwed...

R programming

[tomhath]

There's no point in learning R if you don't have a solid background in statistics.

Re:

[gweihir]

You can still produce a lot of nonsensical results. Like people that learn Java or JavaScript but have no solid ComSci background produce a lot of bad software.

Re:R programming

[umafuckit]

I disagree. By learning R I developed a much better grounding in statistics than I had previously. This is because it doesn't hold your hand and you have to actually think in order to use it.

Re:

[Anonymous Coward]

With so much "fluff" in your typical college diploma... Why not divert all the people...

If you aspire to be a code monkey for the rest of your life, don't go to college. Just try to convince a prospective employer that your awesome C++ skill is just what they need; maybe it is, this year.

Re:

[twistedcubic]

Frankly, R shouldn't be your first language, and probably not even your second.

R is not a general purpose programming language, even though you can write R programs. There are R programming courses (Johns Hopkins, Coursera), but you can get pretty far with using R as a statistical calculator, without ever writing a complete program (ala Octave or Matlab).

Really???

[Anonymous Coward]

" IBM... ...investing heavily in training programs"

Forgive my cynicism, but that doesn't sound like the IBM I know.

Re:

[tomhath]

IBM sells consulting. They have highly trained experts available in whatever technology is hot this year, just write the check and all your problems will go away.

Headlines should rhyme, all of the time.

[mark-t]

It's catchy and bright, if one's not uptight,
To catch up on news with a poetical muse.

No?

(crickets)

(goes away and sulks in corner)