This is a really interesting and quite relevant discussion for me since my son just got out of college with his masters and entered the job market. He really likes “big data” but he is not a math quant/super smart math or computer science or stats student. So, I asked him a bunch of parenty questions like “are you sure you are set up for jobs that have the term data scientist or data analyst in the job title?” I was worried he was focusing on too much of a niche. I was concerned he was ignoring other jobs based on more traditional skill sets for his degree.
I won’t go into a long story about the thing but he just got a job in Atlanta at a very good starting salary and he has already found a place to live up there with people he knows. About half the students in his grad class/cohort have found jobs, the rest are still looking. I thought he was going to be looking for a good long time but nope he walked in, in the 8th inning, team down by 3, and hit one out of the park, as if he knew all along it would happen. It is weird because he is still a kid in my eyes.
Oh, and I totally agree that AI will virtually replace the human element in much of the data sci jobs. I doubt seriously my son considers that at all which is yet another thing I worry about.
This is what I do for a living, and I have a slightly different view about this from some others above.
To me understanding the mathematics is required but not remotely sufficient. Instead, I think of data as something akin to a reluctant witness. There is a true story the data can tell, even wants to tell, but you have to know how to ask it. Ask incorrectly, or too much too soon, and you can get nothing. Even worse the data can be tortured into giving you the story you want to hear, rather than the truth.
Getting the truth is more art than science. But when you find the truth in terms of a new insight, it is very mentally rewarding.
I think to be a good data scientist is more than knowing math, stats, data, or programming. One needs to know about the business they are doing analytics for. I talk to the analysts about the insight they are deriving from Big Data should be actionable. To know whether an insight is actionable one needs to know about the business. It is one of the biggest challenges I find when looking for a good data scientist.
This is a career that D is considering. She is a Applied Math/CS double major without real direction as a rising junior. While looking at 2018 summer internships, data scientist positions kept popping up. She is leaning towards doing an internship to determine if its a good fit. Her choice of study abroad programs is a software engineering program that admits math and computer science majors (so not necessarily engineering only). Is software engineering also a component of data science careers?
@NEPatsGirl One of S2’s friends has the same math/CS combination, also a junior, and has interned with a Seattle start-up in artificial intelligence.
On another topic, the friend’s father has a PhD in Bioinformatics, a data-heavy, math-heavy computer job dealing with writing programs for new drug discovery. Most of his college cohort bailed on their core education and made millions on Wall Street instead.
Although CS and software engineers have a lot of overlapping skills, the focus is different. Many software jobs will be open to either.
Interesting. My daughter is in the second year of a bio PhD program, and is now taking a class in bio data science. She’s learning to code in some specific language used in bio data analysis. She’s techy enough to do well in it, and savvy enough to appreciate the value of it, but she’ll be the first to admit coding isn’t particularly “fun.”
My daughter has a joint econ/math degree and is working as an analyst. These jobs are very popular now at ecommerce companies (like Amazon.) I think students with degrees in many areas - econ, math, statistics, CS, supply chain, etc. are qualified for these positions (based upon the job postings I saw that she applied for- they usually said a degree in a quantitative area).
I don’t think these jobs are going away soon - as others have said it is the interpretation of the results of crunching data or figuring out what is significant or which data to look at or how the data reflects what is going on in that business that is important to succeed as an analyst.
I think AI is the difference between a Data Scientist and a traditional statistician, currently. A good data scientist would have AI expertise, a statistician would not. It is the traditional statisticians and data analysts who may get replaced with the new generation of data scientists who have a broader skill set.
It already feels like data science has peaked and people are moving on.
A few years ago I got a certificate in data science from Johns Hopkins via Coursera, and mostly what I learned was that data science was a lot less interesting and harder to apply than I’d hoped. Still, I signed up to get automated emails listing data science jobs in the San Francisco/Silicon Valley area. It definitely feels like there’s less of a demand for data scientists than there was a year or two ago. Other jobs requiring knowledge of AI and machine learning do seem to be increasing, though. Fortunately, much of what people learn in data science is applicable to AI and machine learning.
@oldfort “I think to be a good data scientist is more than knowing math, stats, data, or programming. One needs to know about the business they are doing analytics for.”
Can you take a new grad who is highly skilled in math, stats and CS (both programming and data structures etc.) and teach them about the specific business that you want them to apply those skills to? It would seem like the good ones already understand the value of actionable information and just need to learn more about the specific industry.
Yes, of course, and that is the type of person I like to hire. The caveat is that the person is actually interested in solving problems, rather than just being in love with the math for math’s sake.
ETA: But the people that most catch my eye are applicants that have taken the initiative to try and do a project in my field. That alone puts someone far ahead of someone who doesn’t take that effort. The work doesn’t have to be perfect, but the approach and decisions made serve as a a great discussion topic during the interview.
My Alma Mater, the University of Rochester, has a new data science institute and growing program. I was one of the first statistics grads there, decades ago, and if I were there now, I’d jump at the opportunity for a degree that could be applied in so many areas.
It seems to me that having solid skills in CS, math, and stats provides great career flexibility even if “data science” happens to fall in popularity in a few years. You can do many things with those skills.
