2013 breakthrough technology "small mobile phone Big data"

The information gathered from the cheap phones can be analyzed to help scientists understand the day-to-day behavior and travel of people, and even to understand the spread of disease.

An epidemiologist at Harvard Public Health School's office, Caroline Barclay, pointing to a telephone signal tower on the map of Kenya on the computer screen, the data collected from the cell phone towers played a key role in the fight against malaria.

While studying the data, she and her colleagues found that people were 16 times times more likely to send text messages and phone calls near the towers of the Kericho, and 3 times times more to send text messages and phone calls in one part of Africa's north-east Victoria, L. than in other regions, These areas have been identified by the health sector in the relevant countries as a malaria transmission hotspot. In these areas, mosquitoes are frantically spreading malaria. By analyzing the data of the signal towers, they charted the route of malaria transmission, which means it is clear that "the suffering of a journalist is bound to cause more people to develop malaria." ”

Caroline is currently building a new predictive model that includes this path map. Finding the source of the disease is not difficult, and more difficult is how to use the signal towers and mobile communications company data to analyze when the sick patients come and go. Caroline said: "Although some people are engaged in immigration, patient medical registration, but in Africa, such data are no different." So in order to solve the spread of the epidemic, we think of a way that no one has ever used before. ”

Data mining, from mobile communications enterprises to obtain the owner of the use of records, and with government departments issued disease Prevention Bulletin. "We cannot take into account all malaria patients, but we can warn and prompt people according to the route of malaria transmission," he said. ”

Caroline and her husband, Nansen Igo, wrote a research paper on 15 million mobile phone users in Kenya last year in the journal Science. One is an epidemiological expert, a mobile data expert, two people in Africa for 18 months of research, through mobile phone data on the ethnic division, disease transmission and other projects carried out, their findings not only unique insights, but also can effectively help people have a healthy life.

Igo said: "This is the future of epidemiology." If you want to exterminate malaria, you have to use this method. ”

A new page

There are 6 billion mobile phones around the world, which produce countless data every day, including location, business activity information, search records and social records. Among them, there are 5 billion mobile phones distributed in developing countries, and the vast majority of mobile phones are cheap, function simple "broken" mobile phones, can only call, send text messages. But the function of the mobile phone does not affect its "big data" function, through the signal tower tracking, can roughly describe a person's life trajectory, which is not only useful for the study of epidemiological transmission, but also can be used to study employment trends, social pressure, poverty, transportation and so on.

For many poor countries, the data collected through cheap mobile phones is more useful than detailed, real-time but very little information. "In developing countries, it is impossible to carry out effective censuses, to get detailed traffic information, and the government does not even have data collection equipment," said Alex Pentland, director of the Institute of Human Dynamics at MIT. But there is one thing that can be seen everywhere, that is the mobile phone. In the past few years, the rapid popularity of mobile phones, even the developing countries are almost one. "Data analysis software and methods have been popularized, clearing the way for data mining."

The case for efficient use of mobile data was the 2010 Haiti earthquake. The earthquake in Haiti in January 2010, killing at least 200,000 people, researchers based on data from Haiti's largest communications operator, analyzed 42 days before the quake and 158 days after the quake, and found that 630,000 people fled the capital, Port-au-Prince, within three weeks of the quake, And the number of people who went to the cholera outbreak was calculated 12 hours after the data were obtained. Because the operator's data is more detailed, can almost do real-time monitoring. More crucially, the researchers used the data to develop a disaster prediction model that found that when the undersea people were in disaster, they tended to take refuge in places they had visited.

First-present scale

Last year, Orange, the famous mobile operator, released a mobile phone call record for 5 months in Ivory Coast of Africa, containing 2.5 billion messages, a gold mine for data experts. Nearly 100 research institutes around the world are doing research with these data, and mobile data mining takes shape. At the same time, the use of data for business and privacy protection is also beginning to emerge. However, in order to maximize the effectiveness of data mining tools, it is not enough to have phone records, and more detailed data, such as small-scale questionnaires.

Caroline hopes to eliminate malaria completely by digging up cellphone data, "This is the future of epidemiology, and if we want to exterminate malaria, we have to use that approach." ”

(Responsible editor: The good of the Legacy)