Estimating high-resolution daily encounter networks with activity-travel diaries
Résumé
This paper presents a novel method for estimating high-resolution daily encounter networks using activity-travel diaries, specifically tailored to the densely populated Île-de-France region. The motivation behind this study is the profound impact of the SARS-CoV-2 pandemic and the critical need for understanding human contact patterns for better epidemiological forecasting and intervention decision-making. The methodology employed involves constructing a behavioral synthetic population of Île-de-France using multiple open data sources, leading to the development of a detailed activity-based model. This model is instrumental in deriving large-scale multi-layer contact networks and estimating age-stratified contact matrices, focusing on contacts within household and overall layers. The approach uses activity-travel diaries to ascertain the spatiotemporal co-presence of individuals, thereby deducing potential and effective contact networks through systematic pairwise comparisons and a Bernoulli model for contact probability. The results highlight significant insights into daily encounter patterns, exhibiting power law distributions with cutoffs in location size distributions and daily contact numbers. Furthermore, the study reveals distinct contact patterns across different age groups in various settings like household, work, education, and leisure. This research, while initially aimed at transport and mobility, demonstrates its significant potential in modeling infectious disease transmission, offering a comprehensive view of human mixing patterns and their implications for public health strategies and interventions.