Urban Mobility Flows from Mobile Phone Data - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2014

Urban Mobility Flows from Mobile Phone Data


Understanding human movements is critical for different scientific domains. In order to deploy efficient networking solutions, a clear view of human mobility patterns is required. The same applies for urban planning, where the global mobility flows can determine the optimal deployment of infrastructure. Human mobility also plays a major role when analyzing the ways diseases can spread in a population. Significant research efforts have been conducted in this direction, aiming at understanding how people move as a first step, and proposing models of such mobility as a second step. Recently, as people are more and more connected, network traces have received particular attention as a source of information about human mobility at large scales. However, previous studies have focused on developed countries, and whether the observed patterns and models are applicable to developing countries remains an open question, due to differences in the lifestyle, country's infrastructure and modes of transportation. Indeed, a clear understanding of human movements would be crucial for the progress of such countries, especially in highly populated urban regions where new transportation infrastructures are being deployed. In this work, we explore Call Detail Records (CDR) of Orange customers in Abidjan, the economic capital of Ivory Coast. The dataset, made available within the context of the D4D Challenge, provides the position of each caller – approximated as the base station's location – at every time he/she initiates a call or sends an SMS. We start by analyzing the temporal, spatial and geographical characteristics of the calls, which allows us to capture differences between distinct times of the day and different days of the week over multiple geographical regions of the city. We propose a method to distinguish between typical and outlying behaviors in the CDR dataset, enabling the detection of special events such as the New Year's Eve and football games played during the Africa Cup of Nations. Our approach also allows us to infer which moments can be aggregated in order to characterize macroscopic mobility flows that provide a view of the globaland local mobility flows in Abidjan, as well as of their daily evolution.
Fichier principal
Vignette du fichier
UrbanModelling.pdf (446.56 Ko) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-01095558 , version 1 (15-12-2014)


  • HAL Id : hal-01095558 , version 1


Diala Naboulsi, Marco Fiore, Razvan Stanica. Urban Mobility Flows from Mobile Phone Data. UM - Urban Modelling Symposium, Oct 2014, Lyon, France. ⟨hal-01095558⟩
255 View
146 Download


Gmail Facebook X LinkedIn More