Mapping a City's Rhythm
A phone application highlights hot spots and will soon show where different urban “tribes” gather.
Over the course of any day, people congregate around different parts of a city. In the morning hours, workers commute downtown, while at lunchtime and in the evening, people disperse to eateries and bars.
While this sort of behavior is common knowledge, it hasn’t been visible to the average person. Sense Networks, a startup based in New York, is now trying to bring this side of a city to life. Using cell-phone and taxi GPS data, the startup’s software produces a heat map that shows activity at hot spots across a city. Currently, the service, called Citysense, only works in San Francisco, but it will launch in New York in the next few months.
On Wednesday, at the O’Reilly Emerging Technologies conference in San Jose, CA, Tony Jebara, chief scientist for Sense Networks and a professor at Columbia University, detailed plans of a forthcoming update to Citysense that shows not only where people are gathering in real time, but where people with similar behavioral patterns–students, tourists, or businesspeople, for instance–are congregating. A user downloads Citysense to her phone to view the map and can choose whether or not to allow the application to track her own location.
The idea, says Jebara, is that a person could travel to a new city, launch Citysense on her phone, and instantly get a feel for which neighborhoods she might want to spend the evening visiting. This information could also help her filter restaurant or bar suggestions from online recommendation services like Yelp. Equally important, from the company’s business perspective, advertisers would have a better idea of where and when to advertise to certain groups of people.
Citysense, which has access to four million GPS sensors, currently offers simple statistics about a city, says Jebara. It shows, for instance, whether the overall activity in the city is above or below normal (Sense Networks’ GPS data indicates that activity in San Francisco is down 34 percent since October) or whether a particular part of town has more or less activity than usual. But the next version of the software, due out in a couple of months, will help users dig more deeply into this data. It will reveal the movement of people with certain behavior patterns.
“It’s like Facebook, but without the self-reporting,” Jebara says, meaning that a user doesn’t need to actively update her profile. “We want an honest social network where you’re connected to someone because you colocate.”
In other words, if you live in San Francisco and go to Starbucks at 4 P.M. a couple of times a week, you probably have some similarities with someone in New York who also visits Starbucks at around the same time. Knowing where a person in New York goes to dinner on a Friday night could help a visitor to the city make a better restaurant choice, Jebara says.
As smart phones with GPS sensors become more popular, companies and researchers have clamored to make sense of all the data that this can reveal. Sense Networks is a part of a research trend known as reality mining, pioneered by Alex Pentland of MIT, who is a cofounder of Sense Networks. Another example of reality mining is a research project at Intel that uses cell phones to determine whether a person is the hub of a social network or at the periphery, based on her tone of voice and the amount of time she talks.
Jebara is aware that the idea of tracking people’s movements makes some people uncomfortable, but he insists that the data used is stripped of all identifying information. In addition, anyone who uses Citysense must first agree to let the system log her position. A user can also, at any time, delete her data from the Sense Networks database, Jebara says.
Part of Sense Networks’ business plan involves providing GPS data about city activity to advertisers, Jebara says. But again, this does not mean revealing an individual’s whereabouts–just where certain types of people congregate and when. For instance, Sense Networks’ data-analysis algorithms may show that a particular demographic heads to bars downtown between 6 and 9 P.M. on weekdays. Advertisers could then tailor ads on a billboard screen to that specific crowd.
So far, Jebara says, Sense Networks has categorized 20 types, or “tribes,” of people in cities, including “young and edgy,” “business traveler,” “weekend mole,” and “homebody.” These tribes are determined using three types of data: a person’s “flow,” or movements around a city; publicly available data concerning the company addresses in a city; and demographic data collected by the U.S. Census Bureau. If a person spends the evening in a certain neighborhood, it’s more likely that she lives in that neighborhood and shares some of its demographic traits.
By analyzing these types of data, engineers at Sense Networks can determine the probability that a user will visit a certain type of location, like a coffee shop, at any time. Within a couple of weeks, says Jebara, the matrix provides a reliable probability of the type of place–not the exact place or location–that a person will be at any given hour in a week. The probability is constantly updated, but in general, says Jebara, most people’s behavior does not vary dramatically from day to day.
Sense Networks is exploring what GPS data can reveal about behavior, says Eric Paulos, a professor of computer science at Carnegie Mellon. “It’s interesting to see things like this, [something] that was just research a few years ago, coming to the market,” he adds. Paulos says it will be important to make sure that people are aware of what data is being used and how, but he predicts that more and more companies are going to find ways to make use of the digital bread crumbs we leave behind. “It’s going to happen,” he says.
The AI revolution is here. Will you lead or follow?
Join us at EmTech Digital 2019.