Hilary Mason is chief scientist at Bitly and cofounder of hackNY, which works to cultivate the next generation of hackers who will stoke innovation in New York. Leslie Bradshaw is COO of the startup Guide, which uses text-to-speech and avatar technology to create animated newscasts from online news, blogs, and social streams. Prior to that, she co-founded the digital agency Jess3, which specializes in data visualization. They met at Bitly’s New York offices to talk about the proper marriage of data and creativity.
Mason: I don't think most people have a really solid definition of what big data is. I prefer the technical definition: You need a computing cluster to store or analyze it. Then there's this expectation starting to creep into the media: "Big data will solve all our problems! We don't need doctors. We'll know how to cure cancer." That has caused this dangerous idea that we don't need creativity anymore. As someone who works with data intensely, I find that scary. Data will tell you whether to pick A or B, but it will never tell you what A and B are in the first place.
Bradshaw: Hal Varian, Google's chief economist, recently said, "The most important job in the next 10 years will be data scientist or statistician."
Mason: He actually said "sexiest," which is even better. Now is an incredibly exciting time to be working with data. For one thing, we actually have data. A lot of the work we're doing on social data is re-creating the past 50 years of social-science research in a large-scale quantitative way. You know those closed-network experiments, like surveying every high school student about whom they've slept with? We can do that by looking at the hundreds of millions of people on social networks.
Bradshaw: In social media, the sample size is humongous. You have the opportunity to see such diversity. Advertisers are using it to their advantage. I'd like to see other people access that.
Mason: This is all art. It's studying human behavior and telling stories about it. How do we change the way people are thinking about some element of social communication? We do that by building data visualizations and telling stories that can then change the common perception of some aspect of how people communicate online.
Bradshaw: The art is in preparing the content for optimal human consumption. The data doesn't just talk back to you. You collect, you analyze, you tell stories. Think of an iceberg. Underneath the waterline are data storage and analysis. Those are your engineers and scientists. Up above is the interface. It's both literal and narrative. It starts with the hard sciences--the math, the analytics--but it ends up with the softest: how to tell the story. We've hired graffiti artists to do data visualization, but first we explained to them the integrity of the data. That coupling is really such an important relationship.
Mason: This is what makes a data scientist rather than a statistician: someone who has the statistics and math background and the engineering capability to make something out of the data, but also the communications ability. I've now hired three physicists, because physicists are used to taking incredibly complicated systems and simplifying them in ways people can understand. I've never hired pure mathematicians, because they're obsessed with the beauty of pure math and not interested in the messiness of reality.
Bradshaw: I want to see the world embrace one phrase: data literacy. There's a scientific element to it. There's storytelling. There's practicality and application to it.
Mason: Not everyone is going to be a data scientist, but everyone should know enough to bring it into what they're doing. Most companies aren't set up to work this way, and most educational institutions aren't set up to teach this way. That's going to have to change.
[Illustration by La Tigre]