"If you want to find a Data Scientist, find yourself a disgruntled postdoc toiling away on brilliant scientific research, but failing to land a professorship because … all the professor jobs are taken!"
That quote is more than just humorous, it points out one compelling answer to the question in the title of the post -- Perhaps all scientists are data scientists, you simply have to lure them into a new domain of study.
Possibly, i have certainly made that point before (and it remains my most upvoted post, so certainly other phd students and postdocs agree with me).
I do think that people need to stop expecting to get a physicist, statistician, economist or applied mathematician (every graduate analysis job I've seen had that wonderful qualification) as most of those people already have really well paying jobs in finance (or satisfying careers in academia), and open their eyes to the fact that for many data science roles, social scientists are probably a better fit.
If you're dealing with numbers generated by people's interaction with a website, a handy background to have is in some form of quantitative social science. (I am of course horribly biased, by being a quantitatively trained social scientist).
In any case, I expect data science to go through a dot.com like boom and then a horrible crash, so it may be a good idea to get the skills (and possibly qualifications) now, while the sun still shines and people are still hypnotised by the promise of big data, rather than the tedium and slog of extracting value from it.
The thing to note is that data scientist interviewing culture at companies is a function of who their first data scientist is or who the founder or whatever things the first data scientist should look like. This is why you have a wide variance from companies that look at people only with applied math and a background in Clojure to companies that require data scientists to be well versed in Java to companies that require statisticians. It has reached the stage where I think it is easier to sell yourself as an awesome scaling/data engineer and quietly do whatever data analytics you want when you get in.
FWIW, I agree with you that quantitative social science is a great background to have. My data science team lead has a background in Neuroscience and he seems to be having fun applying that intuition to recommendations.
To the reply below (for some reason I can't reply directly to you, which is weird).
Yes, I agree that on average, social science people tend to like maths less. However, you're not looking for the average social scientist, you're looking for the ones who weren't satisfied with SPSS (like me). The other, more common kind, are not likely to end up in a data scientist position (though perhaps this may change).
Hilariously enough, after I learned R, SPSS made way more sense to me, while before that it just seemed way too easy to generate reports without a shred of insight.
My wife works in marketing in Research and Analytics. Hard science people do the math (including creating the particular profiles/techniques) much better, on average, than the social science people. While they may have used menus of SPSS a bit, they hit a wall when they hit the scripting language, or at least learn at a glacial pace from the smaller sample sets I've seen.
Yeah, another social science guy here. My employer has found my statistical and behavioral knowledge very valuable, but I'm not much of a coder yet so it doesn't seem like I'm yet able to market myself as a data scientist. I've been doing some serious self-directed training in Python and trying to get better at R (moving from SPSS and SAS)
most of those people already have really well paying jobs in finance (or satisfying careers in academia)
I've always taken the data scientist distinction to be the startup world's answer to the "quant" designation. Quant jobs are a lot better than "just a" programmer jobs, even at hot startups, so I see the "data scientist" title as an answer to that. It's a startup quant.
Yeah, I've actually heard people use the terms interchangeably. Although 'quant' is almost exclusively used in the finance world. I interviewed for a couple of those jobs and got caught up on stuff I've never heard of before, like pricing bonds and combining interest rates.
That quote is more than just humorous, it points out one compelling answer to the question in the title of the post -- Perhaps all scientists are data scientists, you simply have to lure them into a new domain of study.