École Nationale Supérieure de
Techniques Avancées

Résumé

In recent years, the mobile phone data automatically collected by network operators has shown great promise in describing various aspects of human activity. In addition, it can be provided with a short delay and at low cost. These data are therefore of particular interest to companies and institutions that need to adapt quickly to changes in people’s behaviour. In particular, electricity suppliers such as EDF are interested in tracking the evolution of behaviours associated with high energy consumption. Here, we show that anonymised information derived from Orange subscriber data is effective for tracking the evolution of presence in the French departments. In particular, we define an index of presence at work and show that it accurately captures the dynamics of workplace attendance. Since these data are delivered with a few days’ delay, we compare the performance of short-term forecasting models to identify the most accurate one. We believe that this work is a step forward in the integration of mobile phone data for operational situations.