This story is over 5 years old.

How Mobile Phones Can Track Census Data in Hard-to-Reach Places

Knowing where people are is essential to tracking disease and distributing resources.
November 17, 2014, 2:30pm
Image: orangecrush/Shutterstock

Anonymous mobile phone records were recently used by a team of researchers to map out population density in Portugal and France. This proved surprisingly accurate, and now the methods could be used in African countries to improve censuses and even track the spread of malaria.

Researchers from the University of Southampton, Université Catholique de Louvain, and the Université Libre de Bruxelles, published a study in the journal PNAS that claimed population mapping using anonymous mobile phone records could be of great benefit to low-income countries. They worked with Flowminder, a nonprofit advocating data use in response to disasters, and population mapping project Worldpop.

Phone towers receive and log call data, allowing for researchers to determine the approximate location of a phone user. With millions of users, officials can put together a pretty accurate view of population density. While this study tested the model in France and Portugal, it could be particularly useful in poorer regions where census data is lacking and where the data could help in disease control and the provision of services and infrastructure.

"The proliferation of mobile phones (MPs) offers an unprecedented solution to this data gap," the researchers wrote. The GSM Association forecasts that the number of mobile subscribers in Sub-Saharan Africa will grow 43 percent by 2014.

The researchers behind the study are now moving on to Namibia, which has some of the worst rates of inequality in the world. Andy Tatem, a professor at University of Southampton's Department of Geography and Environment and director of Worldpop, told me in a phone interview that they will collaborate with the malaria control program and the main phone operator to map population changes and malaria risk.

By tracking population density through phones and monitoring disease prevalence, the researchers can observe how both malaria and people are moving, which areas are most at risk, and who needs assistance.

In some countries [a census] has not been conducted for 20 or 30 years

While mobile phone use is clearly growing in the area, it's not abundant, but Tatem said this could be accounted for in the findings. "There are plenty of household surveys being done that can give us information on phone ownership and the biases that can come along in terms of how people are using their phones, which groups are being left out, which regions have ownership than others," he said. "We account for that in the mapping and communicate any uncertainties there."

This can't totally replace traditional census methods, but complements them to fill in the gaps. "In some countries [a census] has not been conducted for 20 or 30 years," said Tatem. Without accurate census data, governments can't efficiently provide services and infrastructure to the largest populations that need it the most.

Different countries throw up unique challenges for enumerators, from inclement terrain and weather to threats of violence. In 2008 in Sudan, for instance, some enumerators had difficulty crossing land close to the Nile due to heavy rain while others were attacked by assailants who set fire to several boxes of documents. Using mobile phone data could offer an alternative.

The researchers are keen to point out that the data collected is anonymous. "All we know is a SIM card was there that made a communication," said Tatem. "We don't know who those people are." This is one reason the practice can't replace traditional censuses completely.

Another challenge, he said, is convincing phone companies and governments that this method is worth pursuing further.

Once Namibia's findings are published, and if successful, the researchers have a better chance of spreading the research into more regions. "Through Flowminder there are various other countries we can work with and test this methodology in other places with perhaps lower phone penetration nets," Tatem said.