Audrey Duval $^{1}$ *, Quentin J Leclerc $^{1,2,3}$ *+, Didier Guillemot $^{1,2}$ , Laura Temime $^{2}$ **, Lulla Opatowski $^{1,3}$ **

$^{1}$ Epidemiology and Modelling of Antimicrobial Resistance, Institut Pasteur, France
$^{2}$ Conservatoire National des Arts et Métiers, France
$^{3}$ Université de Versailles Saint-Quentin-en-Yvelines/INSERM, France

* these authors contributed equally
** these authors contributed equally
+ [email protected]

This repository contains the relevant contact networks and analysis code for the paper "An algorithm to build synthetic temporal contact networks based on close-proximity interactions data".

Please note that this project makes use of data previously collected as part of the i-Bird study. Full details are available in "Detailed Contact Data and the Dissemination of Staphylococcus aureus in Hospitals" (Obadia et al, PLOS Computational Biology 2015). For the version of this repository which contains the data, please refer to: https://gitlab.pasteur.fr/qleclerc/network_algorithm

Each script in this folder is named according to the figure it generates. The helper_functions.R script contains two utility functions used to analyse the contact networks.

This folder contains all the figures and supplementary figures of the paper.

This folder contains the relevant observed and synthetic contact networks generated using our algorithm (note: although we generated 100 synthetic networks for each example we describe in the paper, here we only include 10 of each for data storage purposes).

This folder contains the observed contact network, the patient admission and discharge data, and the staff schedule data (please refer to https://gitlab.pasteur.fr/qleclerc/network_algorithm for those files). It also contains an Excel file used to aggregate staff categories into groups.

This folder contains contact networks generated using our algorithm. This includes full reconstructed networks (SimulatedCtc), reconstructed networks with observation bias (testRecord), full random networks (RandomGraph), random networks with observation bias (RandomGraph2), and re-simulated full networks (ReSimulatedCtc).

This folder contains full networks generated for sensitivity with the recurring contact probability set to either 0.1, 0.5 or 0.9.

This folder contains full networks generated for sensitivity using either one or two weeks of data only.