by Noam Cohen, The New York Times, November 8, 2009
DOES Twitter have a T.M.I. problem? And, no, I don’t just mean the Twitter users who share too much information about their lives, social, medical or otherwise.
Simply put, there is way too much information on Twitter — lately, it defies navigation. In January, there were 2.4 million tweets a day, according to Alessio Signorini, a researcher at the University of Iowa. By October, he reports, there were 26 million tweets a day.
Why should we care about information overload at Twitter? Isn’t Twitter about the individual experiences — a Tweeter and her followers — not the totality of millions of Tweeters around the world?
Perhaps this is true for most users. But the promise of Twitter — the reason Google and Microsoft have paid to be able to search millions of Tweets — is that it gives the best approximation of the pulse of the world: How popular is the new iPhone? Did Kanye West make a spectacle of himself at an awards show? Or, more ominously, what is it like when there is a shooter loose on an Army post?
Until lately, the main way to make sense of an urgent outpouring of tweets on a particular subject was to use text searches: look for the phrase “Fort Hood,” for example, or maybe an agreed-upon label, “#fthood,” within tweets. Yet during events like the shootings on Thursday at Fort Hood that left 13 people dead, this method is useless. Hundreds of “relevant” tweets pop up every minute, most repeating the same news reports over and over again or expressing concern from far away.
Which is why a new feature that Twitter says it could unveil in the next few weeks — “geolocation” — holds such potential to make the Twitter rapids navigable.
The idea is to take advantage of global positioning systems on cellphones to allow Twitter users to include a precise location with each tweet. Users would be able, right off the bat, to limit their searches to tweets from a particular location.
“Proximity can be this proxy for relevance,” said Ryan Sarver, the director of the Twitter platform, who led a “fairly small team” of programmers who after a few months are close to completing the geolocation project. “We are about delivering the right information to the right people.”
Even now, before the geolocation feature has been released to developers, visitors to Twitter.com have been able to limit searches by location based on the profiles that Twitter users provide when they sign up.
That simple filter made a huge difference in what a visitor to Twitter’s search engine discovered about the Fort Hood shootings. After limiting searches to those from within 15 miles of Killeen, Tex., a town near the Army post, you easily find messages sent by soldiers describing what it is like to be on lockdown or worrying about their children at school on the post.
The tweets from one, RicoRossi, can suddenly leap to the top, even while Fort Hood remains locked down — “a soldier i treated here said he was waiting in line @ SRP when another soldier stood up and started shooting,” one early tweet read, using an acronym for the Soldiers Readiness Program, where the attacks occurred.
In response to questions from his Twitter followers, he quickly replied with “idont want 2 b 2 graphic so ill stop there, he was there … it was like something out of a movie he said im paraphrasing of course.”
Another Twitter user, JKsTinkerbell76, described the scene at Scott and White Hospital in nearby Temple, Tex., during her lunch break (around midnight late Thursday): “Still on Lockdown. Police and Security everywhere.”
And DaTriggerMan, Killeen, Tex, tweeted Thursday afternoon, “We just got cell phone services back here on ft hood! And I am good thanks for all the concern!!!” And then, four minutes later, “what happen cuz I aint been in front of a tv? We had some shooters on post we good now tho!”
Improvements like geolocation have the potential to make the Internet suddenly relevant to society as it is lived, not just relevant to what happens online. Mr. Sarver imagines features like “local trending topics,” a list of subjects popular in a particular area; or searches for happy hour in a neighborhood of Dallas that will intelligently link tweets about happy hours to the place they were sent from.
Because GPS will provide the ability to become very “granular” with locations, you could mimic through Twitter the banter at the local diner or a barbershop, by limiting a search of tweets to a two-block radius.
There is also the fear of loss of privacy and loss of security as once-local chats become globally public. That is why Mr. Sarver said Twitter would require two “opt in” decisions — at the profile level and again through the application.
For the technological optimists, the cures for information overload, in essence, are better filters and greater context. The more you know about a message — who sent it and why — the better you understand it.
The open-source project Ushahidi, which takes its name from the Swahili word for “testimony,” was quick to understand this. The software allows text messages to be mapped by time and location. It was developed to track reports of ethnic violence in Kenya in 2008. Suddenly mere words can create a moving picture of where violence started and where it intensified.
More recently, through a program called stopstockouts.org, the Ushahidi software tracks a range of medicines for shortages across Africa.
Creating navigation tools for digital information is the next big challenge, said Erik Hersman, a co-founder of Ushahidi who has been in contact with Mr. Sarver’s team at Twitter.
“We have more update-type of information,” he said. “The stream is getting wider and wider.”