Behavioral-epidemic models for COVID-19

This repository contains code and data to reproduce the results of the paper "Comparative Evaluation of Behavioral Epidemic Models Using COVID-19 Data" by Gozzi et al, 2025.

Data

The data used in the paper is available in the data folder. In particular, the data folder contains 9 subfolders, each containing the data for a different region considered in the paper (Bogotá, Chicago, Gauteng, Jakarta, London, Madrid, New York, Rio de Janeiro, and Santiago de Chile). Each subfolder contains the following folders/files:

• contact_matrix: contains a contact_matrix.npz file, which contains the contact matrix for the region. Contact matrices are obtained from Mistry et al, 2021.
• epi_data: contains a epi_data.csv file, which contains the epidemiological data for the region, with daily new cases and deaths.
• google-mobility-report: contains a google_mobility_data.csv file, which contains the Google mobility data for the region, from the Google Mobility Report.
• population-data: contains a pop_data_Nk.csv file, which contains the population in different age groups (0-9,10-19,20-24,25-29,30-39,40-49,50-59,60-69,70-79,80+) for the region.
• hemisphere: contains a hemisphere.csv file, which contains the hemisphere of the region (needed for seasonal forcing).

The source of the data is the following:

Region	Demographic Data Source	Epidemiological Data Source
Bogotá	Observatorio de Salud de Bogotá, Población de Bogotá	Gov.co Datos Abiertos, Casos positivos de COVID-19 en Colombia
Chicago	Census Reporter, ACS 2022 1-year, Total Population	Chicago Data Portal, Daily Chicago COVID-19 Cases, Deaths, and Hospitalizations - Historical
Gauteng	Coronavirus COVID-19 (2019-nCoV) Data Repository for South Africa, Provincial projection by sex and age	Coronavirus COVID-19 (2019-nCoV) Data Repository for South Africa
Jakarta	Population by Age Group and Sex in DKI Jakarta Province, 2020	Daily Update Data Agregat Covid-19 Jakarta
London	Office for National Statistics, Estimates of the population for the UK, England, Wales, Scotland, and Northern Ireland	Coronavirus (COVID-19) Weekly Update, Greater London Authority (GLA)
Madrid	Instituto Nacional de Estadistica, Población por comunidades, edad (grupos quinquenales), Españoles/Extranjeros, Sexo y Año	Ministerio de Sanidad, COVID-19 Deaths
New York	United States Census Bureau, Age and Sex	NYC Health COVID-19 Data
Rio de Janeiro	Instituto Brasileiro de Geografia e Estatística, Population Projection	Ministério da Saúde, Coronavirus Brazil
Santiago de Chile	Instituto Nacional de Estadisticas, Proyecciones de población	Departamento de Estadísticas e Información de Salud, COVID-19 Open Data

Code

The code used to reproduce the results is available in the models folder. In particular, the models folder contains a file for each of the three behavioral epidemic models considered in the paper:

• mobility_model_age.py: contains the code for the Data-Driven Behavioral (DDB) Model.
• compartment_model_age_deaths.py: contains the code for the Compartmental Behavioral Feedback (CBF) Model.
• function_model_age_deaths.py: contains the code for the Effective Force of Infection Behavioral Feedback (EFB) Model.

Additionally, the file utils.py contains the functions used to calibrate the models via Approximate Bayesian Computation (ABC), while the file constants.py contains the values of the fixed parameters used in the models.

We provide an example of how to run the models in the example.ipynb notebook.

Citation

To reference our work, please use the following citation:

@article{
    doi:10.1073/pnas.2421993122,
    author = {Nicolò Gozzi  and Nicola Perra  and Alessandro Vespignani },
    title = {Comparative evaluation of behavioral epidemic models using COVID-19 data},
    journal = {Proceedings of the National Academy of Sciences},
    volume = {122},
    number = {24},
    pages = {e2421993122},
    year = {2025},
    doi = {10.1073/pnas.2421993122},
    URL = {https://www.pnas.org/doi/abs/10.1073/pnas.2421993122},
    eprint = {https://www.pnas.org/doi/pdf/10.1073/pnas.2421993122}
}