Dataminr builds a Twitter-powered early warning system

Dataminr, a New York-based start-up that has been quietly building a global sensor network powered by Twitter, is introducing its technology to the public Monday, showing how its real-time engine can act as an early warning system for enterprise and government customers. Dataminr has signed a partnership with Twitter to access its firehose and is analyzing it in dozens of ways, helping surface breaking news, critical information or highlighting trends from emerging events.
Dataminr combs through 340 million daily tweets on Twitter and its algorithms quickly seize on abnormal and actionable signals that can be analyzed and confirmed as a relevant event for a client. This could be anything from an assassination or general instability in certain countries to government sanctions, natural disasters or on-the-ground chatter about products or trends. Dataminr uses available Twitter metadata along with other contextual factors such as historical and concurrent data to create a mathematical signature for an event, ultimately deciding on the fly whether an event is valuable for decision-making purposes. For example, Dataminr’s clients were alerted 20 minutes ahead of mainstream news coverage of Osama Bin Laden’s death.
Dataminr, which is built heavily on Hadoop and is hosted on Amazon’s EC2, processes 50 terabytes of data per month. The company works to make the data easy to use so clients can find what they’re looking for quickly, Dataminr CEO and founder Ted Bailey told me in an interview. He said previous technologies that tried to look at Internet data relied on things like sources and the authority of sources or keyword frequency. Dataminr is very focused on the now and deriving instant value from the flow of tweets online.
Predicting the now
“It’s not just that we capture early information, but also where the eyes of the world are pointing. That’s a valuable indicator of what’s happening in the world and where the world will focus in the future,” said Bailey. “We have event detection software that is able to pinpoint specific events going on in the world. Instead of predicting the future, we’re very much predicting the present and giving people better understanding of what’s happening right now. And that has enormous value.”
One of the Dataminr’s key strengths is the way it can absorb and merge third-party information or a client’s proprietary data to help add more context and inform Dataminr’s work. That ensures that a customer can harvest unique insights that are geared toward their interests or bolstered by their own data.
Customers can access Dataminr’s API and also view information through a desktop dashboard. Typically, a client will ask questions of the Twitter data or look for specific events to base decisions on. But Dataminr can also surface other information relevant to clients based on their general interests and needs.
Founded in 2009, Dataminr has signed up a number of big banks and hedge funds and has been making inroads with government clients. It’s now planning to roll out its services to a lot more customers  in the second half of the year. The company closed a $7 million round of financing last fall from a number of angels and private investors from the tech and finance industries.
A big fan in Twitter
Twitter has been a big fan of the service and has talked up Dataminr at a couple of events — they got a mention from CEO Dick Costolo in October. Robert Weeks, a spokesman for Twitter, told me by email that Twitter sees a lot of value in what Dataminr is doing, helping the enterprise market be aware of and capitalize on moments that matter. It’s also an example of the kind of start-ups that Twitter likes to encourage, showing off the overall power of the Twitter fire hose in areas like finance and government.
“Dataminr is a strong team building compelling products to help clients grow their businesses. Twitter is committed to working with partners like Dataminr who add value to the hundreds of millions of Tweets sent every day by finding the right signals for the right audiences,” said Weeks.
There’s a growing opportunity in making sense of the flood of data flowing through services like Twitter. More and more, there are valuable insights that can be gleaned from real-time social data if it can be identified and confirmed as critical events for markets, industries and governments. Dataminr has got a good sense of how to leverage that data and make it useful to this growing customer base, who sees more insights emerging from the flood of social data.
Thumbnail image courtesy of Flickr user bettina n