Is it possible to locate the position of a home through the analysis of mobile phone data? This was achieved by researchers from the Institute of Information Science and Technology of the National Research Council (Cnr-Isti), the Universidad del Desarrollo in Santiago de Chile, the Telefonica Chile operator, the University of Turin and the Foundation ISI of Turin. Together they validated for the first time a large set of algorithms, already existing or specially developed, on the telephone records of 65 volunteers in Santiago de Chile. The study, published in EPJ Data Science, made it possible to identify the most accurate algorithm, discovering that telephone calls at night identify the home of the number bearer with greater precision and that a fraction of less than 20% of telephone calls is sufficient for identification.
“Determining the home of an individual through the analysis of his telephone records is an experiment that allows you to increase the ability to automate statistical processes of great importance, such as the estimation and forecast of population density, migration flows, ‘air pollution and the construction of mathematical models for epidemiological predictions, also with respect to the spread of COVID-19 “, explains Luca Pappalardo of the CNR-Isti. “Despite the importance of this detection, the lack of the set of telephone records associated with the area of residence (ground truth data), did not allow us to properly validate the methods currently used, so we were not sure of the accuracy of their algorithms. Our study finally sheds light on the accuracy of these algorithms, while at the same time allowing us to calculate the minimum amount of data necessary for a given purpose of public interest. “
.