BY Rich Griset07 March 2022, 15:01
A student uses his laptop on the Dartmouth College campus in Hanover, New Hampshire, USA, as seen in October 2021. (Photographer: Bing Guan—Bloomberg/Getty Images)
Data science has been one of the fastest growing job fields, with demand for data scientists increasing by 650% between 2012 and 2017 alone according to LinkedIn. Demand should stay that way; from 2020 to 2030, the employment prospects in the field is expected to grow by 22%, according to the US Bureau of Labor Statistics.
To meet this demand, colleges and universities have launched new Masters in Data Science programs. In fact, Howard University, University of Connecticut, and University of California, San Diego have all announced new programs in recent months.
With so many changes happening so fast, what does the future hold? Fortune spoke with data scientists to learn how data science programs are likely to evolve.
Stay up to date with technology
The field has transformed even in the nearly seven years since Grant Long earned his master’s degree in data science at New York University.
“It’s changed tremendously,” says Long, now head of data and machine learning at Flow code, a startup that aims to connect the offline and online worlds through QR codes, mobile-focused landing pages, and other digital tools. “Technology continues to evolve. I think a lot of data science resources are organized around open source packages that have become increasingly popular in various companies. Obviously [at] technology companies where many of these open source projects come from, but also [at] more traditional businesses.
Long says adopting this new technology has been “transformational” for many companies. While programming models like MapReduce were taught when Long was at NYU, master’s students are now more likely to be trained on open-source programs like Apache Spark.
While technology may change, the fundamentals of data science remain the same, says Joel Sokol, director of the interdisciplinary master’s degree in analytical science at Georgia Tech and professor in the H. Milton Stewart School of Industrial and Systems Engineering at Georgia Tech. ‘university.
“In some sense, there’s not a lot of distance between the starting point and the point where you’re already cutting edge,” like natural language processing and deep learning, Sokol says. “The most important thing schools try to do in these programs is try to keep up with what’s cutting edge, because that’s where a lot of professional practice ends.”
With so many new developments in the field, Sokol says it’s common to see professors from other schools take online courses from Georgia Tech, Carnegie Mellon University and Massachusetts Institute of Technology to stay current with latest advances.
As for the impact of artificial intelligence on the field of data science, Sokol warns that what some call artificial intelligence today is essentially more powerful machine learning algorithms, not something like HAL. 9000 from “2001: A Space Odyssey”. Yet, he says, these algorithms will open new doors to data science, potentially aiding the ability to make predictions.
How data science programs focus on practice versus theory
In the field of data science, there has already been some bifurcation between practitioners who focus on the practical side and those who focus on the theoretical side. Sokol says that since there are not many doctorates. data science programs, there will always be more specialization at the graduate level.
“You’ll end up with some master’s programs that are designed for practitioners, where they focus on applying these methods and how to use them, and then some master’s programs will be basically pre-doctoral. programs, where you will learn more about the theory,” says Sokol.
Additionally, Sokol says there are certification options in data science and analytics. In these non-master programs, participants learn the basics of programs and software, but little about the theories and approaches that underlie them. These “boot camp” options will continue to see increased demand for some time as there are not enough people in masters programs to fill the current need for data science and analytics positions; However, Sokol says that as more people enter graduate and undergraduate programs that prepare them for these roles, these boot camps will see less demand.
With a career that includes stints at the Federal Reserve Bank of New York, Capital One and Zillow, Long says he’s witnessed the rise in specialization firsthand. While at Capital One in the mid-2010s, Long says data science was seen as a broader term and the role of a data engineer was just beginning to emerge. Back then, data scientists were doing a lot of data engineering work.
Today, when Long starts up, there are five different titles among the approximately 85 full-time data science employees: Machine Learning Engineers, Analytics Engineers, Data Analysts, Product Analysts, and Data Science Engineers. data.
Long says data science programs of the future will specialize more in computer science and statistics at a technically rigorous level, as well as in analytics. With the growth of packaged data science solutions that require little coding, like Alteryx and AutoML, non-data scientists can perform simple data analysis, but Long says there will always be a strong demand for people who can conduct more complex methods of data capture and analysis. This need will continue to be observed in data science graduate programs.
“Going forward, there will be more program specialization,” Long says. “He will also continue to be stronger [rewards] to programs that emphasize the fundamentals of math and computer science, because these things are not easy to understand.
Because technology is changing so rapidly, employers are looking for people with a solid background in handling data and rewarding them with higher pay rates. In fact, some students graduating this spring are already receiving job offers of $125,000 or more. Programs that emphasize this foundational education will see higher-paid graduates, secure higher-level jobs, and see greater demand from program applicants, Long says.
A focus on real-world experience
A master’s degree in data science helps signal that a job candidate knows the basics, but Long says potential hires are more impressive when they combine that degree with real-world experience.
“What I look for when I see candidates is that they have already been exposed to data in a given role, but they have also had exposure to fundamental theories, computer science and mathematics,” says Long. “It helps signal that you’ve had that foundational exposure.”
Although open-source packages have made some elements of the data scientist job easier, Long says it’s still essential to know the basics.
“In a place like Flowcode, the most important thing is to have people trained to adapt, which is why some of the more classical elements of computer science and mathematics are actually the most important,” says Long.
Find out how the schools you’re considering landed in Fortune’s ranking of top business analytics programs, data science programs and part-time, executive, full-time, and online MBA programs.