Michael Driscoll concentrated in Government at Harvard because “back in 1999, only geeks majored in Computer Science, or that’s what we thought,” he says. “For me, the appeal of government was about impact. It took me a while to realize that information technology was another place where you could have a broad impact.”

Now that he’s gotten over his aversion to geekiness, he’s the cofounder of Metamarkets, a company whose platform provides predictive analytics for digital media companies. He also started the Bay Area R Users Group, which only had ten members when he started it two years ago but has now grown to over 1,200 members who meet up to talk about the R programming language for statistical computing, data analysis, and visualization.

His initial exposure to big data was through the life sciences and bioinformatics. His summer internships and first post-college job was at the National Institutes of Health, where he worked for the National Heart, Lung, and Blood Institute and then on the Human Genome Project. “I worked with genome data, which consists of two things—the sequence, and then the annotation that goes on top of the sequence,” he says.

In the midst of his work at the NIH—during a transition period between departments—he started an online T-shirt retail company, CustomInk, because he wanted to start a company and saw an opportunity when he remembered trying to order shirts for various clubs in college.

After the NIH, he moved to France for a year to work for a company called Genset, helping them do analysis for large-scale genome data. He came back, spent another year at NIH doing bioinformatics, then got his Ph.D. in Bioinformatics and Systems Biology at Boston University.

After he finished his thesis, which focused on microbial genome sequences, he started looking around for a biotechnology startup opportunity.

“But as I looked around at the biotech space, I realized that biotech startups are generally very capital intensive and the return horizons very long, and I’d seen a lot of friends of mine try to start biotech companies that hadn’t been as successful as they’d hoped,” he says. “The other part of it was, a lot of the work being done in the life sciences ought to be done in the academy, not so much in the private sectors. Maybe in a few decades, some of the things people are excited about will start to have commercial implications, but we’re not there yet.”

So he decided to move to Silicon Valley. “This is where the heart of innovation and information technology is,” he says. “I’d spent the last five years working with very large data sets in the life sciences, but I felt strongly that the same set of tools I was using had applicability to other verticals, so I came out here and started a consultancy called Dataspora, which was basically a way for me to sell my skills.” He ran Dataspora for 2.5 years, growing it to several other consultants, focusing on taking large-scale data sets and driving insights out of them for businesses. He recently sold Dataspora to a Boston-based company called Via Sciences.

“The common thread through online retailing, bioinformatics, the Ph.D., then online media markets has been how can we use data to drive value?” he says. “How can we take information and make better businesses, better decisions, and better products?”

He co-founded Metamarkets over a year ago, when he realized how much data is in online media markets. “It appealed to me because I thought, if we could bring statistics, math, and machine learning to bear in verticals outside of the life sciences, there’d be a lot of value to unlock,” he says. “It was this incredibly data-rich vertical that was struggling with some pain points around managing, analyzing, and visualizing that data. It struck me as a huge opportunity.” They raised $2.5 million with IA Ventures leading the round, and raised another $6 million two months ago.

There are three parts to the Metamarkets platform—a data processing layer, an analytics layer, and then a visualization dashboard generated by D3, which is a JavaScript framework that allows dynamic real-time visualization.

“The world is changing in the way we look at webpages,” he says. “They’re no longer these static shipments given to us. We’re embracing that with our data visualization framework, whereby as events happen on the backend, we don’t have to wait hours or days for that to get surfaced to a decisionmaker. We can get it within seconds, with events and graphics updated in real time.”

“Information technology is becoming more and more commoditized, and a lot of the edge information technology companies have is going to move away from the management of data and towards its analysis,” he says. “That’s where algorithms and machine learning become really important. Algorithms are going to be the heart of competitive advantage for technology companies in the next couple of decades.”