Data scientists hold the Rosetta Stone to translate data into dollars, and companies are willing to pay big bucks...
to get them on board, but these new hybrid scientist-software engineers are difficult to come by.
Data scientists typically have an innate curiosity, along with a unique blend of knowledge and capabilities that includes skills in analytics, data mining, machine learning and statistics, plus experience with algorithms and coding. These "unicorns," as they've been called, are also business-minded and can turn their findings into products or services that help companies gain a competitive edge.
This mix of skills sets them apart from other data analysts, said Paddy Hannon, CTO and senior vice president of audience and analytics at Edmunds.com Inc., a car shopping and research website that employees a team of 40 data scientists.
"They know how to integrate analytical work with software engineering," Hannon said. "They know SQL, Python and R; they don't do a lot of work in Excel -- not that there's anything wrong with Excel. But they are fluid in terms of their ability to turn analysis into a piece of living, breathing software."
With data coming at companies from every direction, the demand for those skills is higher than ever -- and companies are willing to pay for them.
The median annual salary for data scientists is $116,840, according to job-listing provider Glassdoor Inc., which places data scientist at the top of its "25 Best Jobs in America" list for 2016. That's up from No. 9 last year, when the median salary came in at $97,835. The best jobs list is based on the career website's "Job Score," a rating determined by the number of job openings, median salary and career opportunities.
A search for full-time data scientist jobs posted on Glassdoor over the past three weeks yielded 2,339 listings from companies, including Edmunds.com, Facebook, Fidelity Bank, Kohl's, Pinterest and Uber.
Data scientist job postings are up 50% year over year on Dice, a job-listing website and resource for technology professionals. On any given day, there are 450 job postings for data scientists -- 230 directly in the job title, a company spokesperson said.
This is a massive increase over the past five years. Dice began keeping track of data science as a role in 2011, when there were only eight data scientist job postings on the site.
A few listings on Dice include Samsung's opening for a data scientist in San Jose, Calif.; Amazon's opening for a data scientist in Seattle; and Kohl's needing a senior data scientist in California. For those early in their careers, Cox Enterprises has an opening for a data scientist intern this summer in Austin, Texas.
The data scientist supply-and-demand problem
Intuit Inc., a financial management software provider, is among the companies with openings for data science jobs. It already employs over 200 data professionals across engineering and analytics, including 30 data scientists. Over the next year, the company plans to increase the number of data professionals by 15% year over year, said Sarah Peterson, communications manager at Intuit. But finding the right people to fill data science roles may prove difficult.
"The combination of what makes a great data scientist is very rare," Peterson said. "The best data scientists have a unique combination of skills. ... And they apply these skills with design-thinking methods and deep customer empathy to proactively uncover critical business insights, recommend product enhancements and identify innovative new business opportunities."
Edmunds.com also has a data science job opening to grow its team. The search for a person with the right amalgamation of skills can take six months to a year, but there have been times when the right candidate came along more quickly, Hannon said. To find talented data science experts, companies like Edmunds.com partner with nearby universities and sponsor events, such as data festivals, to get in front of future graduates.
Paddy HannonCTO and SVP of audience and analytics at Edmunds.com
"You can come right out of school and make good money [as a data scientist], and there isn't slack in demand," Hannon said. "We have all of this data, and we need people who can live it, breathe it and find ways to act on the data."
Demand for data scientists really began to take hold in early 2014, when large companies with mature enterprise data warehouses began to pull unstructured data from previously untapped sources, said Matt Mueller, president of CBIG Recruiting and Staffing, an IT resource management company that focuses on business intelligence and big data analytics. Companies needed experts to investigate data and deliver predictive business results based on their analysis, he said.
Although there appears to be more data science jobs than there are qualified job candidates, the supply will increase as universities add data science programs at the undergraduate and graduate levels. However, that supply will be absorbed quickly, Mueller said.
"More and more companies will need to bring on data scientists, and it is a job and skill set that really will be hard to commoditize," Mueller said. "I still believe demand will outpace supply when it comes to companies requiring data scientists."
Some companies have taken the supply problem into their own hands. Intuit developed an eight-week boot camp for individuals who are either in a Ph.D. program or recently completed one.
"What we've found is that Ph.D.'s from various fields generally have a good handle on working with large data sets and have a strong research orientation, which are qualities you need in a good data scientist," Peterson said. "What they oftentimes lack is the ability to apply these skills in a business setting."
To remedy that, Intuit's two-hour, once-a-week boot camp, Intro to Data Science, helps candidates for data science jobs apply their research skills to a specific business or product, and creates a candidate pool Intuit wouldn't otherwise have been in touch with, she said.
Data science in action
One of the many brainchildren of data scientists at Edmunds.com includes the CarCode platform, which gives car shoppers a way to communicate with dealers via text message. Development is ongoing, as data scientists work to improve the tool; they are analyzing scenarios where people might not get a response and which topics result in fewer responses. The answer may be an automated way to respond to general questions about dealership hours and directions, Edmunds.com CTO Hannon said.
Data science experts with the car shopping website also work to crystalize its recommendation engine. They analyze behaviors to predict interest in things such as motor vehicle safety features, or to identify someone who may be interested in a lease, rather than a car loan. The site's predictive model is updated every 60 seconds to provide a personalized Web experience, according to Hannon.
Another is Edmunds.com's True Market Value tool -- developed a decade ago, before data scientists were called data scientists -- to help consumers negotiate fair prices at dealerships.
"[Data scientist] is an interesting role, because what they do goes beyond analysis; they impact products in a real way," Hannon said.
That's the case at Intuit as well, where data scientists use their analytical and engineering skills to solve complex customer problems. One example is QuickBooks Financing.
Small businesses need access to loans to grow their business, but many are denied because they put all their small business expenses on a credit card, which hurts their FICO score. QuickBooks data shows that they pay their invoices on time and what their cash flow looks like. Intuit uses that data, with their permission, to match them to qualified lenders, Peterson said. To date, $275 million worth of loans have been facilitated for Intuit's small businesses customers, she said.
Data scientist shortage hampers IoT efforts
Strategies for hiring data scientists
How to think like a data scientist