Bitly scientist searches for online data patterns

When a restaurant opened near her New York City apartment, Hilary Mason found its menu uninspired: salmon, arugula and all the other usual culinary keywords without any of the risky dishes that make for great dining.

“I thought, ‘This must be the median West Village restaurant; it’s just so boringly average,’ ” she said.

Most people would have written a tepid Yelp review and left it at that. Mason looked to the data for truth. The 33-year-old wrote a program to crawl the Web and download menus from New York eating places. It took her down a rabbit hole of restaurant exploration. She didn’t figure out the perfectly average spot, but she learned that there are 173 different burgers to order in the West Village —but 363 in the East Village, and at lower prices.

Mason, a self-described nerd, calls this “menu hacking,” and it’s one of her many geeky side projects. Her day job is chief scientist at Bitly. The startup is best known as a link-shortening service, a way of making obnoxiously long URLS more compact for sharing on Twitter or anywhere else. This year, Bitly is introducing data products for professionals developed in part by Mason and her team of six scientists and engineers. One, dubbed Bitly Realtime, tracks terms that receive sudden bursts of attention. Another is a reputationmonitoring system. The goal of the products is “to give people a Spidey sense about what’s going on on the Internet that’s relevant to them,” Mason said.

Plenty of other companies promise to cut through the noise of the social Web: Mass Relevance and Dataminr license data from Twitter and other sources and sift it for meaning. Mason says Bitly is different, because it doesn’t just track what’s being shared but also records which links people actually click on.

That’s “valuable information that today is underanalyzed,” said Nick Gall, a vice president at researcher Gartner.

Bitly is not Mason’s first startup experience. She studied computer science at Brown and started teaching at a small university in Providence in 2004. While there, she researched virtual worlds such as “Second Life” and formed a company to track how users behave in them. She sold it for “beer money” a few years later.

“If I knew then what I know now, there would have been another zero on the end,” she said. In 2008, she moved back to her hometown of New York and joined Path 101, which analyzed data gleaned from resumes to help people figure out the right career steps to their dream jobs.

The company “failed miserably” in 2009, she said. “It turns out it’s important to build a product and not just a bunch of data models.”