The temporal dynamics of group interactions in higher-order social networks

This repository contains the code required to reproduce the results presented in the following paper:

• I. Iacopini, M. Karsai & A. Barrat (2023), The temporal dynamics of group interactions in higher-order social networks , Nature Communications 15, 7391 (2024).

Data

This study relies on publicly available datasets that have been collected and released in previous publications:

• CNS University interactions collected by the Copenhagen Network Study, presented in Sapiezynski et al. 2019 and downloaded from here;
• DyLNet Preschool interactions presented in Dai et al. 2022 and downloaded from here;
• Conferences interactions collected by the SocioPatterns Collaboration, presented in Génois et al. 2023 and downloaded ---upon request--- from here.

Original data are saved into the data-raw folder.

Structure

• code contains most of the Python scripts
• data-raw contains the sub-folders CNS, DyLNet, and Confs where the original data are stored
• data-curation contains the Jupyter notebooks that perform the pre-processing of the raw data from data-raw, saving the results in data-processed
• data-processed contains the data after the pre-processing described in data-curation
• data-analysis contains the Jupyter notebooks that analyse the empirical data already pre-processed in the data-processed folder
• model contains the Jupyter notebooks that run the model and analyse the synthetic data

Dependencies

The code of the paper strongly relies on a number of exeternal Python libraries. The code has been originally run on a machine containing the following Python dependencies:

• matplotlib 3.5.1
• networkx 2.8
• numpy 1.21.6
• palettable 3.3.0
• pandas 1.5.3
• powerlaw 1.5
• scipy 1.10.0
• xgi 0.5.6