Mining mobile-phone data for the public good.
Nathan Eagle, a research fellow at the Santa Fe Institute in New Mexico, believes that mobile phones offer more than a way to communicate. In his hands, they can provide windows on the social structure of communities, information that can lead to better public-policy decisions, and unexpected sources of income for people in poor countries.
For years, Eagle has been mining cell-phone data captured by service providers around the world. Using algorithms he developed as a graduate student at MIT, he strips all identifying information from call logs and looks for patterns in where people go and how they use their phones–patterns that can reveal how social networks are affected by outside forces. For instance, he is working with city planners in Kenya and Rwanda to understand how slums grow and change in response to events such as natural disasters and declines in crop prices. And earlier this year, Eagle began using phone-derived data to build a more accurate model of the spread of malaria in Africa. Previous models had relied on spotty information about people’s movements, collected in sporadic surveys. With a better picture of how the disease spreads, governments can improve the policies designed to fight it.
In February, he launched Txteagle, a service that lets any company send cell-phone users simple tasks such as text translation. Participants are paid with credits that can be used for phone service or redeemed for cash at special kiosks. A pilot program in Kenya paid a few cents per task and was too successful for its own good. Within hours of its launch, the ranks of users swelled into the thousands; within days, all the tasks were exhausted.
Eagle plans to relaunch the service later this year in Kenya and other countries, including Rwanda, Indonesia, and the Dominican Republic, with two changes that he hopes will make it sustainable: capping the amount of money a person can make in a day, so that completing tasks becomes more like a hobby than a job, and offering more tasks, such as identifying objects and people in digital pictures or deciphering distorted words from scanned books.
Simple sensors to detect residents’ activities.
Walls can talk, and Shwetak Patel, an assistant professor of computer science and electrical engineering, captures their stories: tales of how people move through their homes and how they use electricity, gas, and water. Patel has shown that each electrical appliance in a house produces a signature in the building’s wiring; plugged into any outlet, a single sensor that picks up electrical variations in the power lines can detect the signal made by every device as it’s turned on or off. This monitoring ability could be particularly useful for elder care, but there was previously no practical way to achieve it, because it would have required numerous expensive sensors.
Last year, Patel did something similar with ventilation systems, designing a sensor that detects subtle changes in air pressure when a person leaves or enters a room. More recently, he’s shown that slight pressure changes in gas lines and water pipes betray the use of specific appliances or fixtures, such as a stove or faucet. Patel believes that providing people with information about their patterns of resource consumption can help them reduce it. He has cofounded a startup that will provide consumers with utility bills itemized by appliance.
Using software to send diverse radio signals.
With Ashoke Ravi’s help, future cell phones and netbooks won’t need separate circuits to transmit multiple radio signals (over a cellular network, Wi-Fi, and WiMax, for example); a single transmitter will handle them all.
Radios that use software to receive signals over different wireless protocols exist already, but progress has lagged on the transmission side. Much of the difficulty has involved building amplifiers that can cope with the different power levels needed to transmit over the varied distances typical of different wireless networks.
Ravi, a researcher at Intel, built a software-controlled transmitter that solves the problem. Instead of changing the power level to transmit different signals, its amplifier can attenuate or boost the outgoing signal by combining the output of two oscillators that operate at a constant power level. His design allows the amplifier to be optimized for a single power level, increasing battery life.
Ravi expects devices incorporating the technology, such as laptops capable of switching seamlessly between 3G and Wi-Fi networks, to be on the shelves within five years.