Much has been said lately about the diversity of skills required by a data scientist to successfully translate data-driven insights into tangible business results. Now, management consulting firm Booz Allen Hamilton is taking this understanding one step further and applying it to an organization’s entire data science team.
“The emerging field of data science is a fundamental piece to unlocking the power hidden within the avalanche of data,” says Booz Allen’s vice president and data scientist Dr. Joshua Sullivan. “Our government and commercial clients see the greatest return when their data science teams possess the expertise to help ask intuitive questions of data, and the ability to translate the newly discovered insight into information that matters to the organization’s core mission.”
The consulting firm has outlined this approach in an infographic exploring the anatomy of a data science team. The key point is that the most successful teams are composed of more than just data scientists. They also include so-called “domain experts,” drawn from the ranks of the industries and professions that are looking to utilize the data, along with innovative but logical thinkers — such as mathematicians, physicists and musicians — who can offer different types of insights and perspectives.
“There’ve been tremendous benefits from integrating pure research folks with computer scientists,” says Jason Escaravage, a principal at the McLean, Virginia-based consulting firm. In particular, he adds, “we’ve learned that when supporting a customer it’s critical to have domain expertise.”
The domain specialist understands the data that exists for a particular field — be it credit reporting, oil and gas exploration or military intelligence — and can address what Escaravage terms the three V’s of data: its volume or quantity; its variety or mix of structured, unstructured and real-time or streaming data; and its velocity, referring to the time between when the data is collected and the decision has to be made.
Another key practice is to rotate the team’s domain specialists in and out of the team every six to nine months, to avoid isolating them from the very industries the in-depth knowledge of which they are bringing to the team.
On the flip side, armed with new insights from their data team stint, the domain experts are in an ideal position to popularize them within the ranks of the business and its management group.
Recognizing that data science is a capability set that will be in high demand for the foreseeable future, Escaravage says Booz Allen is working with several universities, including the University of Maryland, to try and build out the pipeline.
At Maryland, Booz Allen is one of several companies offering internships and future employment opportunities to data scientists in the making. As part of the curriculum, students are given the chance to work on actual problems confronting different industry sectors, developing their analytical skills in a real-world setting.