These data contain information on a caller’s location and spending, so could help researchers better understand people’s movements, social interactions and economic conditions.
The potential applications of Call Detail Records (CDRs) — which are made anonymous to protect privacy — were highlighted by a series of case studies published last month (5 November) by UN Global Pulse, an initiative set up to find ways to use vast sets of data for development.
According to Anoush Tatevossian, spokeswoman for UN Global Pulse, “skimming the research” to highlight examples applicable to development was essential to raise awareness beyond the academic community.
“One of the first steps is making the bridge between researchers and those who could apply [such research findings],” she tells SciDev.Net.
One of the ways CDRs can be used is by analysing the location of populations based on which phone mast their devices connect to.
As well as tracking emergency migration after the Haitian earthquake in 2010 and combating malaria in Kenya and flu in Mexico, this method could also play a role in city planning, the report finds.
“Helping governments plug the gaps in their census data ... is one of the most promising potential applications.”
Miguel Luengo-Oroz, UN Global Pulse
One case study shows how information on the call locations of half a million individuals over a five-month period allowed researchers to outline typical travel pathways through Abidjan, Côte d’Ivoire’s highly congested capital city, thus pinpointing where to concentrate new infrastructure and bus routes.
Other applications combine location with other metrics, such as the pattern of calls an individual makes within their social network, or expenses incurred due to phone use, for insights into socio-economic patterns and census data.
It is this role in helping governments plug the gaps in their census data, particularly in developing nations with scant resources, that is one of the most promising potential applications, says UN Global Pulse’s head of research, Miguel Luengo-Oroz.
But he and Tatevossian warn that the models used in the case studies were developed over a long period, and that the field is still several years from offering tried and tested tools that yield results in a practical timescale.
Proving that the approach yields cost-effective and reliable results that can be generalised across countries, and working out how data use will fit with privacy regulations, are some of the obstacles that still need to be overcome, says Luengo-Oroz.
Mirco Musolesi, who studies large-scale data mining at the UK-based University of Birmingham, agrees that privacy issues are one of the main potential sticking points, but is optimistic that any problems can be overcome.
Furthermore, the general openness of companies to share their data banks with not-for-profit initiatives, along with a growing awareness among the development community of the power of CDRs, means the area is likely to grow in importance in the coming years, he says.
Link to full report