Data Science Back to School: Accelerate Your Education
Are you looking to get a data science degree and join the workforce as a data scientist? Then you’re not alone, as thousands of young people around the world are following that same path with the hope of tapping into the high demand – and large salaries – that come with being a data scientist. As you embark upon your education journey, keep some of these tips in mind to maximize your education and job prospects.
We have been hearing how hot data science is since Harvard Business Review declared data scientist “the sexiest job of the 21st century” way back in 2012. Apparently, not a lot has changed (except perhaps the implosion of Hadoop), as a January report from LinkedIn concluded that data scientist was the number one job for 2019, thanks to a solid average base salary of $130,000 and abundant growth opportunities.
Compared to 2018, the number of openings in data science increased by more than 4,000 jobs in 2019, which is nearly a 60% increase from the previous year, according to LinkedIn. But that number is just the 12-month delta; the number of actual job openings for data scientists is much, much higher. To wit: A year ago, LinkedIn concluded that there were 151,000 unfilled data scientist jobs across the United States, with “acute shortages” in places like San Francisco and New York City, where you can imagine that salaries are much higher.
One of the folks looking to hire data scientists is Obed Louissaint, the vice president of talent at IBM Watson. Louissaint is chasing that pool of candidates in the hopes of making them IBMers.
“In general, everybody is hiring for AI, whether you’re a bank or whether you’re a retailer,” he tells Datanami. “Everyone is looking for that particular skill. So we’re finding the challenge of the market being saturated with people looking at the same supply.”
With upwards of 10,000 resumes submitted to IBM per day, Louissaint and his colleagues in IBM’s human resources department have their work cut out for them. Automation helps IBM weed out the candidates who are obviously not a good fit so the HR professionals can better evaluate who’s left. “Thank God for AI!” he says.
When hiring for data science or AI positions, what technical skills you have are big factors in whether you get the job. Depending on the position that Louissaint is looking to fill, the skills might demand knowledge of languages like Python and SQL, the ability to work with TensorFlow or PyTorch, or familiarity with data visualization tools.
While college degrees can help in these departments, they’re not as important as you may think. “We’re in a world where it’s less about what degrees do you have, as it is what skills do you have,” Louissaint says. “Degrees still play a significant place in overall credentials. But credentials that are going to be differentiating are things like certifications and badges.
The data science degree you are aspiring to will open a lot of doors in data science, but it won’t open all of them, Louissaint says.
“Employees and candidates should be migrating to those particular jobs and the way they qualify for those jobs is by demonstrating their technical expertise and prowess through skills, badges, and credential, not necessarily through degrees,” he says.
Skills are King
One way to demonstrate technical prowess is by enrolling in and completing shorter courses, many of which are online. DataCamp, which has presented data science courses to more than 4 million people, offers statements of accomplishments that people can put on their resume to show proficiency in a certain area.
DataCamp offers more than 300 courses on various topics, ranging from building models in R and Python to using BI tools. “We cater to novice, intermediate, and the person who wants to go very deep in AI and deep learning, the latest technologies out there,” says Martijn Theuwissen, a co-founder of DataCamp.
While college degree programs have their place, DataCamp is focused more on helping students and employees keep their skills relevant to what the market is demanding, Theuwissen says.
“There is a value in having a degree. I don’t want to be dismissive toward traditional education,” he says. “As far as keeping yourself up to date with what’s moving and getting statements of accomplishments around that, that’s where a platform like ours come in, the whole continuous education, continuous training aspect, which is also very appealing to employers.”
About 70% to 80% of DataCamp users are in the workforce, although employers pay for just 20% to 30% of DataCamp subscriptions, according to Theuwissen. With that said, at $300 per year, DataCamp subscriptions are not what you would call expensive, especially compared to today’s college tuitions. Theuwissen also points out that more than 1,000 college professors use DataCamp in their classrooms. The New York City company offers free six-month subscriptions for professors to use in their classes.
Data engineering courses are currently trending at DataCamp. “Compared to two years ago, I would say that data engineering….pops up a lot more,” Theuwissen says. “It’s becoming way more popular in terms of the demand that we see.”
The company’s data engineering curriculum focuses on technologies like Apache Spark, Scala, ETL, and working with cloud services and databases. “It’s more technical work, versus the data science, which is more statistical work,” he says.
While the latest engineering and science skills will always be in demand at DataCamp, Theuwissen says he’s noticed more demand for general data savviness among data analysts and others in general marketing or finance positions. As the democratization of data continues in organizations, data citizens, as he calls them, want to expand and improve their data abilities.
“We see more and more demand for what we call non-coding courses, so courses for managers that sit together with data scientist and want to be dangerous enough to ask the right questions. We see a huge movement toward that,” he says. “The data citizen is a person who’s a data-savvy person. I would describe myself in that category. I know how to do my analytics for my job, but I wouldn’t say I’m a data scientist with a very deep statistical knowledge.”
Soft Skills Matter
Prospective data scientists need to have the technical and statistical chops to get the work. There’s no doubt about that. While the AutoML tools are improving and chipping away at the algorithmic knowledge that data scientists need to keep on tap in their brains, they cannot replace a human data scientist (at least not yet).
But being the smartest person in the room can only get you so far in the real world. In addition, data scientists should consider adding some soft skills to their repertoire. That’s the advice given by Eric Johnson, who was the CIO and an SVP at Talend when Datanami interviewed him last month but who has since moved on from the company.
“As CIO, I’m not usually poking at their tech skills,” Johnson says. “But what I’m most concerned about is can they hold a conversation? Can they influence people? Can they work as a team? Can they get work done that is cross-functional, where maybe not everyone is on the same page? How do they work through that problem and how are they successful in that? To me, that’s what I’m looking for. If they can’t do that, even if their technical skills all check boxes, I think I need to keep looking.
“I don’t want a bunch of folks who are just super technical and super smart, but can’t interact with each other,” he continues. “That smells like disaster for me.”
Being able to spot correlations hidden in huge sets of historical data and then devise algorithms to automatically respond to those deviations in live data is definitely a skill that will not go out of demand any time soon. Don’t be afraid of channeling your inner Sheldon, but try to keep in mind what employers want.
The best data scientists, Johnson argues, will be able to temper that raw intellectual horsepower with a business-focused savvy that helps the company flourish where others just see numbers.
“Back when I was in school, it would have been really helpful if someone would have given me the basics around what are these core business processes and what does your business process look like and use cases around that,” he says. “There are some pretty common business process and business challenges that are… horizontal, like the close-to-cash process.”
It’s not uncommon to find data science courses offered through a business school, although most are hosted within departments of Computer Science and Engineering. One can’t reasonably expect a recent college graduate to have perfected their understanding of a specific line of business. But having a little bit of experience of how business works in the real world can make a profound impact on a young data scientists’ potential to build data solutions for the real world.
“What a great way to get a competitive edge, having that business background,” Johnson says. “For me as an employer, if somebody came to me and said I not only have data science techniques down — the modeling down, understanding the technical components — but I also have an operations background? Oh my god, you’re sort of worth your weight in gold.”
It’s a wonderful time to join the data science trade. Business is booming at the moment, and opportunities appear to be everywhere. The great thing is that everybody gets to craft their own path. Candidates can take many routes into a career in data science. The hope is these tips may help your career take off just a little faster.
Editor’s note: This story has been corrected. DataCamp is based in New York City, not Boston, Massachusetts. Datanami regrets the error.