With a mathematics degree from the U.S. Naval Academy, a master’s in business from Notre Dame, and years of experience flying military helicopters into war zones in the Middle East, Michael Hood already seemed plenty qualified to “help solve the world’s big problems.”
He’d held this lofty goal since childhood in Alaska, where—as a self-described “sci-fi geek”—he devoured books with protagonists using math and engineering wizardry to vanquish bad guys or protect good ones. With a heroic 8-year career as a Navy Aviator nearly behind him in 2015, he envisioned a future as a different kind of hero, developing and commercializing technologies to make life better for people. But during a trip to Silicon Valley during grad school, he got a rude awakening.
“I realized that unless you can really write code, you are functionally useless in a modern tech company,” says Hood, now 32. And as far as his dreams of being a data scientist? “I saw them as the Ninjas of the tech world—these guru hackers who could solve third-order differential equations in their head. It felt totally over my head.”
Fast forward to today, and Hood—a graduate of the 12-week Galvanize Data Science Immersive program—has earned a seat among the ninjas. He’s working as a principal data scientist for consulting firm ProKarma on an ambitious project aimed at helping hospitals prevent the number of patients who must be readmitted. He loves the work, which—in retrospect—was far less mysterious than he’d envisioned. And as he points out to data scientist wanna-bes like he was, there are a lot more jobs like it out there.
“Data is like the new crude oil, bubbling up out of the ground, and it is the data scientist’s job to figure out how to refine it into something useful,” he says. “There is a huge demand for people with these skills and not nearly enough supply to fill it.”
Supply and Demand
According to one study by management consulting firm McKinsey and Company, the number of data science jobs in the United States alone will exceed 490,000 by 2018, but there will be fewer than 200,000 people qualified to fill them. Those who rise to the challenge will be rewarded, with salaries ranging from $140,000 to $232,500, according to a 2014 report by executive recruiting firm Burtch Works. (By contrast, the average annual income for a lawyer is $131,990; doctors earn about $184,000).
Despite such demand, less than one-third of U.S. News & World Report’s Top 100 Global Universities offer degrees in data science. Many MBA programs fail to touch on it. And many graduates, like Hood, aren’t very familiar with it when they get out of school.
“The number of data science jobs in the United States alone will exceed 490,000 by 2018.”
“I don’t see it as a failure of education per se,” says Hood’s Galvanize Instructor Ben Skrainka, who co-founded the Immersive Data Science Program in Seattle to help fill the knowledge gap. “It’s just that in the old days, what everyone learned in business school (like how to do a business analysis with Excel) was enough.”
Today, Skrainka says, the size of the datasets have grown exponentially, as has the computational power available to process them. The modern data scientist’s task is to develop models for turning these vast heaps of incoming data into tangible answers to key questions: Which credit card charge is most likely to be fraudulent? Which patient is most likely to get sick again after being discharged? Which customer is most likely to respond favorably to X advertisement? And wisely using this data has a much bigger payoff than it once did.
“Five years ago there were hardly any data science jobs. Today, it’s like we’re prom queens,” says Skrainka.
A Mid-life Pivot
After honing up on his coding skills on his own, Hood applied to the Galvanize program, struggling until 4 a.m. over a challenging take-home quiz (including statistics, Python programming, and math) as part of a somewhat grueling application process. To his surprise, he got in. And while it was rigorous, the mystique around the job soon began to lift.
“Put simply, we learned math and computer science techniques that allow you to find meaningful patterns in the data,” he says.
Skrainka teaches what’s known as “CRISP-DM”, a standard workflow model that ushers students through a series of key steps—from determining a realistic business objective for the data mining, to asking the right questions, to making sure they have the right data sets, to cleaning up the data, to developing models for finding patterns, to deploying and testing the model. They learn about machine learning, statistical inference, and working with data at scale, are taught new languages like SQL and NoSQL, and are exposed to Agile, a collaborative approach to software development.
Skrainka says Hood, with his passion for learning was just the kind of student Galvanize is intended for.
“He was a pleasure to teach, and to help him make this big career change was really satisfying.”
Hood was hired within five days of completing the program, at a salary twice what he was making before he started. He’s still dreaming up big ideas for how data science can change the world. But now he has the skills to help make it happen.
“Galvanize is a lot like military boot camp,” says Hood, who should know. “You come in as this raw unshaped person and you come out transformed. I’d recommend it to anyone.”