Breaking news: Twitter has a new tool for breaking news.
Along with its new search filters, Twitter announced a plan to beef up another discovery tool today: the micro-blogging site is partnering with CNN and data-mining firm Dataminr to create a new alert system for breaking news.
The system, appropriately enough, will be called Dataminr for News, and it will use the data-gathering company’s proprietary algorithms to pick up on patterns. The company will use its mining process to let newsmakers know when words and phrases are appearing in seemingly meaningful ways, which may help the people who make the news pick up on eyewitness accounts earlier. After all, if the words “giant hamsters,” “earth on fire,” and “hamsters eating our children while world burns” appear enough times, there may be some kind of hamster apocalypse occurring. But many times events are unpredictable enough that there’s no way to set an alert for a particular incident (as in the example I just gave).
Although Twitter partnered with CNN to test this project, it will be a resource for the larger journalist community. Twitter noted that the alerts will roll out to all global news agencies in 2014, though it’s not clear whether this will just include traditional outlets like the AP and The New York Times, or whether blogs and newer outlets would be included. Daily Dot asked Twitter whether smaller online newsrooms would gain access, but have yet to hear back (and are still wondering if we made the cut).
Dataminr explained more about the upcoming function on its website: “Journalists are able to set targeted alerts for certain types of breaking information that are then delivered automatically via application, email, pop-up, even instant message, depending on a user’s preference. Dataminr for News can be customized depending on a user’s particular topics of interest and regions of focus. Dataminr for News can also be directly integrated into existing internal client systems.”
Photo via Flickr/Ben Sutherland
Pure, uncut internet. Straight to your inbox.