Many studies are focused on using data derived from mobile phones to construct statistical indicators. Mobile phone data have the advantage of providing information with both high spatial resolution and at high frequency, allowing applications such as measurements of the spatial or temporal details of population presence. Nonetheless, using mobile phone data to construct statistical indicators raises difficulties: data from a single operator are not representative of the whole population and they often lack socio-demographic detail, which limits their quality for many applications. This article is based on a database of mobile phone records from subscribers collected by a large French operator. It aims to offer a view on the potential, but also the problems posed by mobile phone data, specifically by illustrating how indicators of residential populations can or can not be estimated from them.

Online complements can be accessed on the French version.

Article on one page

To cite this article

Sakarovitch, B., de Bellefon, M.-P., Givord, P. & Vanhoof, M. (2018). Estimating the Residential Population from Mobile Phone Data, an Initial Exploration. Economie et Statistique / Economics and Statistics, 505-506, 109–132.
https://doi.org/10.24187/ecostat.2018.505d.1968