Mobile Phone Data Used to Predict Poverty in Rwanda
Anonymized data generated by mobile phones has been used by researchers to show how wealthy or poor people are.
Personal information that mobile devices gather such as a person’s location often gets anonymized by stripping it of names, home addresses, phone numbers, and other obvious identifying details. Such metadata often get shared, and underlies popular services such as Google’s real-time monitoring of road traffic.
Scientists at the University of Washington in Seattle and the University of California, Berkeley, analyzed data from billions of phone calls and text messages from 1.5 million subscribers to Rwanda’s largest mobile phone network.
The data captured details about individuals such as social networks, travel patterns, and the amount and timing of communications.
The researchers also conducted phone surveys of more than 850 subscribers of the network. The investigators asked respondents questions about what their housing was like, about whether they had access to assets such as motorcycles or electricity, and about other indicators of their wealth or poverty.
The scientists used the asset data and phone usage patterns of these respondents to predict the wealth and poverty of the other 1.5 million or so subscribers. The wealth and poverty maps their system generated agreed with those made using detailed surveys of the Rwandan population conducted in person by the Rwandan government.
Limited resources make censuses and household surveys rare in developing countries such as Rwanda, hampering research into how poverty and wealth are distributed in those nations. For instance, previous research suggests that national statistics on economic production may be off by as much as 50 percent in much of Africa.
The scientists noted their approach could lead to new ways to quickly analyze poverty and wealth at a fraction of the cost of traditional methods, helping uncover the kind of details that are essential to sound economic policy.
Details of the findings can be found in the 27 November issue of the journalScience.[Via Spectrum]