“It would be nice to see things like a graphic of spikes in the conversation, what time they happened … and be able to scroll back to that time to see what happened at that particular moment,” Costolo said. He envisions a tool that would allow users to “follow along with Twitter in a DVR mode” rather than sift through a reverse-chronological stream of tweets.
Twitter is trying to find ways to mine data from the crush of information on its site. Costolo explained that the difficulty lies in filtering “the signal from the noise” during live events, which often cause Twitter to crash from an influx of traffic.
“[The] ability to track and monitor the moments within an event, either as they happen or to catch up with them, is something we want to enhance,” Costolo said. “We want to make that experience even better, curating the moments within the event, the media from it, and making it that much easier to navigate.”