Ellerdale Project: Making sense of Twitter’s ‘fire hose’
Two million times an hour, people around the globe see fit to tweet. Sometimes it’s “I’m taking a shower,” and other times, it’s “I feel an earthquake” explains Ellerdale Project co-founder Arthur van Hoff (left, above with co-founder Jens Christensen). “The former is not very interesting, but if suddenly you see a few hundred earthquake tweets, you probably have some real information. The problem is, how do you extract possibly interesting data from this gigantic flow?”
Enter Ellerdale, a downtown Menlo Park-based technology startup that makes products that help people understand large, real-time data feeds including the Twitter fire hose. The company maintains a demo site trends.ellerdale.com that continuously parses Twitter and the web for the most popular topics. Where Twitter can tell you what keywords or authors are trending, Ellerdale’s state-of-the art semantic processing can tell you what topics (and related sub-topics) are trending in real time. Ellerdale software “understands” relationships (e.g. Justin Bieber and Miley Cyrus are both teenaged singers), which is easy for humans but a very difficult problem for computers. As you can see on trends.ellerdale.com, Ellerdale’s algorithms make it straightforward, in real time, to see trending topics in almost any category – people, politics, science and more.
While semantic processing is not new (Tim Berners-Lee has been evangelizing the Semantic Web for years), the depth (and speed) of Ellerdale’s product is very new – and impressive. Ellerdale is betting that modern marketers, product managers and state-of-the-art cloud applications will find lots of ways to use this information in a future driven less by raw data and more by higher-level inferences drawn from that data. Will Justin and Miley date? Ellerdale may be the first to know.
