Alexandre CEBEILLAC, CNRS, France
Eric DAUDÉ, CNRS, France
Daily mobility is essential in studies of spatial structures and dynamics. They play a role that has not yet been fully assessed in the spread of epidemics at an urban scale, because it is intertwined with other vulnerability factors. This is particularly true for vector-borne diseases (dengue, Zika), for which the level of exposure of an individual to a pathogen varies according to his or her presence in certain places (where the vector is present) at certain times (when the vector is active) (Daudé et al, 2015). To assess the role of mobilities in the spread of a pathogen, we propose in the MO3 project a process-based method, i.e. modeling the main mechanisms that govern human movements. Daily mobilities are calibrated from multisource data and allow the simulation of individual exposure times in risky environments.
This presentation focuses on the description of algorithms for generating mobile synthetics individuals in Bangkok, a metropolis where dengue fever is endemic. The result is the creation of millions of agents with spatialized agendas, i.e. a sequence of activities that can be performed in different locations at different time steps, and that vary according to the place of residence and the socio-demographic characteristics of the agents.Different data sources are mobilized: individual data from Twitter, coupled with land use layers obtained via Google maps, allow us to obtain spatialized agendas from which it is possible to extract travel statistics between two activities performed (Cebeillac et al, 2017-18). This information is completed and adjusted according to the aggregated time use data provided by the Thailand Statistics Center. Finally, the use of hourly attendance indices of places obtained from millions of Facebook check-ins allows us to weight the attractiveness of the different areas of the city. The use of a probabilistic approach and adjustment parameters (weights given to the data) finally allows to generate different mobility scenarios.
Mots clés : daily mobilities|synthetic population|agent-based model
A104990ED