Data Science is a new term and a new job title that has been receiving quite a bit of hype. There have been arguments over the definition of this term, and whether or not it truly describes a new field of endeavor or is just an offshoot of statistics, software programming or business intelligence. Another take is that many of the definitions of a Data Scientist can be met by few if any individuals, but really define a team. The television show Numb3rs ran with new episodes in the USA from 2005 through 2010. In many ways, the show was a prototype for such a Data Science team. Let's look at the roles on the show, and see how they might translate into your organization.
The Mathematician or Statistician
Charlie Eppes, boy genius, who grew into a young professor of Mathematics at the fictional CalSci. Like so many professors, he supplemented his income by consulting. In his case, to the FBI and his brother, applying mathematics and statistics to solving crimes of all sorts. His breadth and depth of knowledge was remarkable; unlikely to be matched in the real world. However, an applied mathematician or statistician, with knowledge of a branch of mathematics or statistics relevant to your problems is essential in either a Data Scientist or a data science team.
Computational Statistician, Computer Scientist or Software Developer
Amita Ramanujan, a student at CalSci who achieves her doctorate in computational mathematics, becomes a professor at CalSci, helps with the consultations to the FBI and, as a side note, dates her one-time thesis advisor, Charlie Eppes. In many ways, Amita is the closest to being a Data Scientist of anyone in the show. Equally adept at mathematics, physics, statistics and hacking, Amita often acquires the required data from disparate sources, and transforms Charlie's mathematical visions into working code. If you hire an Amita, you might just have all you need to get Data Science producing real solutions for you.
Over time, both Charlie and Amita gain a fairly impressive domain knowledge of criminalistics.
Subject Matter Expert (SME)
And speaking of subject matter experts, this is where Numb3rs really prototyped what is required to make Data Science valuable in solving the crimes, er, problems at your organization. There were many SMEs, both as regular characters, and as special roles for specific shows. The regular characters included the FBI agents, from Don Eppes, the lead agent, and Charlie's brother, to his team, with David Sinclair and Colby Granger surviving the entire series. By the end of the series, these two FBI agents were taking turns suggesting and explaining mathematical approaches. Other FBI agents who were on the team for one or more seasons include Terry Lake, Megan Reaves, Liz Warner and Nikki Betancourt. Megan was also a profiler.
Also of important note, was the Eppes brothers' father, Alan Eppes. Alan was a retired city planner for Los Angeles. His knowledge of building, regulations, and city's byways, processes, neighborhoods and interactions, were often instrumental in understanding the results of Charlie and Amita's calculations.
Another regular was Larry Feinhardt, holding the Walter T. Merrick chair at CalSci. It may surprise you to learn that a theoretical physicist and cosmologist, interested in studying the heavens, string theory and zero point energy was a crucial member of a crime solving data science team, but he was, other than a brief hiatus aboard the International Space Station.
More important than these regular cast members however were the guest SMEs. There were some recurring roles, such as the drop-out with a knack for baseball stats, or the mechanical engineer who was more interested in how things failed than in how to build them. Flame propagation, biology, disease control, cryptology, cognition, gaming, chemistry, forensic accounting and more specialists all are needed at one time or another to solve the crime.
Retrospective
For me, the lesson is that while you may find an individual who can be creative with the right math or statistics, find, extract and massage data, have sufficient domain expertise and write sophisticated code, turning their creative algorithms into real solutions, you're more likely to need a team. And even that team will need additional help, from within the organization or outside consultants. Your regular team and "guest stars" mathematicians, frequentists, Bayesians, engineers, scientists, accountants, business analysts, and others to bring the best decisions out of your data.
And if you want to check out how realistic the mathematics was in Numb3rs, check out The Math of Numb3rs from Cornell University. For more on the cast, characters and show, look for the CBS official Numb3rs site, and, of course, the Wikipedia article on Numb3rs.
Update: 20120722: I'm honored by the mentions and retweets on Twitter from those who love the Numb3rs analogy. Thank you all.
.
