By Foued Melakessou, University of Luxembourg
Collecting, analyzing and modeling the distribution of the mobility flows is an important factor that enables the optimization of the current transportation infrastructure (vehicle routes, bus and train schedules, etc.) in order to efficiently support the user demand. The analysis of Call Detail Records has captured the attention of traffic and transportation researchers to optimize people’s mobility. We have used Call Detail Record (CDR) datasets, provided by Orange Senegal , combined with demographic data and maps (base stations geolocalization), in order to compute, characterize and identify the mobility flows of Senegalese population. Studies and researches were performed using mobile communication data made available by SONATEL and Orange within the D4D Challenge. Scilab has been the central component of our model analysis platform. Thus we have extracted realistic human mobility models adapted to the Senegal use case. Daily traffic demand profile of each mobile phone base station has been modeled, by considering voice and also messaging activities. The evaluation of mobility models will help to better design and develop future infrastructures in order to better support the actual demand. This facilitates the impact evaluation of efficiency of the transportation services and infrastructure modifications, such as the addition of new roads. We have built mobility maps from CDRs’ statistical analysis, based on the Scilab NARVAL toolbox . Mobility patterns have been highlighted, in respect with specific vizualization tools. A classification of each base station has been performed into urban, suburban and rural modes. An algorithm has also been developed in order to detect traffic anomalies, based on the computed daily profiles. The second contribution corresponds to the generation of interantenna and inter-district mobility graphs for each month of 2013. This work was supported by the research project named MAMBA  that intends to propose and validate a multimodal mobility platform that relies on new Internet technologies to interconnect different mobile services to provide relevant travel advice based on the context of the users, so as to optimize overall system performance.
This Data was made available by ORANGE / SONATEL within the framework of the D4D Challenge. The author would like to thank Orange and SONATEL for the availability of these CDR’s datasets.
Big Data, CDR, Mobility, Graph