Finding the Right Path for Your Data Science Education
When it comes to learning data science, there’s no shortage of options available. You can pick up a book, read articles on the Web, or participate in a massively open online course (MOOC). Or, if you desire a more structured approach, you can take a six-month boot camps or even apply to a two-year degree programs. So which one should you choose?
Like most things in life, there are no hard and fast rules, and what works well for one person may not work at all for another. More specifically, the right path for you may depend on where you are in your career right now, and where you’d like to go.
For example, if you’re young and looking forward to a long career in data science, it might make sense to pursue a Master’s degree from one of the many two-year data science graduate programs that are offered by universities across the country. Having a solid foundation in math, statistics, and machine learning will prepare you for a bright future in data science, no matter what industry you wind up working in.
But if you’re in the middle of your career as a business analyst, for example, and are seeking to move up the analytics ladder into the data science realm, you may be better off boning up on your skills through a MOOC, or participating in one of the many data science bootcamps that have sprung up over the past few years. The bootcamps, in particular, will top you off with specific job skills that are applicable to your profession, and the certificate that comes with it will help your resume stand out.
No single path works for everybody–it’s different strokes for different folks, says PK Agarwal, who’s the Regional Dean and CEO of Northeastern University-Silicon Valley, which currently offers data science bootcamps and will be launching a two-year data science degree program.
“The people who come to a master’s program come for different reasons,” Agarwal tells Datanami. “It’s not just pure knowledge acquisition. They take a longer view of life, and if they move into management or a CXO position, they’ll need the master’s degree.”
But if your goal is increasing your data science competency to get the job done now, then the shorter bootcamps may suffice, says Agarwal, who was the CTO for the State of California under Governor Schwarzenegger.
“My instinct is we’ll see a lot more traction in the certificate programs,” he says. “The largest amount of traction is going to be in the short six-to-eight month certificate program, where they can really upskill these people.”
At the recent Strata + Hadoop World show in San Jose, Northeastern University’s Silicon Valley branch conducted a survey to try to quantify attendees’ thoughts and feelings regarding the state of data science education. Some of the findings were not especially groundbreaking, such as that data science and analytics were the most coveted skills, and that most of the attendees were already working in the data science field.
But one survey finding stood out: More than 60 percent of those surveyed said they did not feel they had the necessary skills to be a data scientist, even though they are currently working in the field of data science. This glaring statistics reflects several things, including the incredibly fast evolution of big data technologies. But what it perhaps reflects the most is that persistent data science gap, which appears to be widening as demand for data scientists outstrips supply.
A Hybrid Approach
Datanami covered many aspects of the data science gap with a special three-part series in late March. Graduate programs, bootcamps, and MOOCs are all expected to play a role in the solution, with different approaches fitting better with different individuals and their needs.
But according to Agarwal, the best educational results will come not from just doing one thing or the other. Rather, a hybrid approach that combines both in-classroom instruction and online education appears to be the best approach.
“We don’t have the time to be totally face-to-face all the time. But I don’t want to be sitting at my desk and trying to make sense of this. I need something in the middle,” he says. “We’re a big believer that, on the education side, things are shifting from pure online to a hybrid mode.
Northeastern University’s Strata + Hadoop World survey found that attendees were highly responsive to a hybrid educational approach. “They’re pretty unanimous in saying that we’d like this in a hybrid format, a combination of face-to-face and online,” he says.
That’s the approach that the university, which has 28,000 students across four main campuses, uses with its bootcamp style courses. It starts out largely online, with a physical get-together every two weeks.
The hybrid keeps travel costs of face-to-face meetings low while still providing enough human interaction to solidify the topics in students’ minds. “When you do self-learning, you think you’ve got it, but it’s all [academic] until start to apply these things,” Agarwal says. “There are a lot of studies that say the learning outcome is significantly better when you have some face time. But the rote learning part, pretty much we can do online.”
Northeastern isn’t the only university espousing the benefits of a hybrid approach. Indiana University Bloomington offers a hybrid option for its Master’s of Science in Data Science program, whereby the student can complete the first year’s courses online, and then complete the second year’s courses on campus. Syracuse University, Rutgers, and Columbia University also offer data science course in a blended mode, while many other colleges offer classes either online or on campus, but not both.
You’ll find many routes to obtaining your data science education. And thanks to evolving educational methods, you’ll hopefully find success sooner than later.