Resumen
Currently, the widespread penetration of mobile devices makes it possible to use the data collected during their operation to analyze traffic flows. Associating mobile phones with their owners, it is possible to judge by the movements of mobile devices, for example, the structure of traffic flows in the city. These tasks have nothing to do with some kind of tracking of mobile subscribers, only anonymized data is used here. Moreover, due to the specifics of the tasks of analyzing traffic flows, individual movements are not fundamentally interesting here, but it is the general (aggregated) data on movements between the selected objects (sections) that are needed. This article presents the main directions of using aggregated data of cellular operators, such as searching for changes in traffic flows corresponding to important social events, measuring these changes, searching for novelty in traffic flows, clustering areas and connections between them, as well as assessing the distribution of people across city districts. The article describes an approach to identifying anomalies in traffic data (anomalies in the time domain are investigated), which is based on the analysis of aggregated data of cellular operators. The paper presents a computational experiment that demonstrates the correctness of the proposed method.