SEIR PDE estimates effective distance based on time series data of infection, enabling us to understand infection-related human mobility networks.

Tetsuya Yamada* and Shoi Shi*, Estimating infection-related human mobility networks based on time series data of COVID-19 infection in Japan. bioRxiv.

This project requires the following libraries.

• NumPy
• SciPy
• Pandas
• odeintw > 0.1.0
• emcee > 3.0.0
• corner > 2.2.0
• tqdm

All source codes used in our manuscript are in this folder.

• models.py:
  • The SEIR model expressed by ordinary differential equations (ODE)
  • The diffusion model expressed by partial differential equations (PDE).
  • The diffusion model that was used for estimating impacts of the effective distance on the scale of the pandemic.
  • The diffusion model in inter-prefecture network graph.
• mcmc.py:
  Run Markov chain Monte Carlo (MCMC) using affine invariant methods to estimate parameters, judge convergence based on an auto-correlation function, and visualize a result using estimated parameters.
• cartogram.py
  Distort a map based on the effective distance from a reference point (e.g., Tokyo) and local connectivity such as geographical distance between nearby prefectures.

You can use command line interface to run source codes in src.

All data used in our manuscript are in this folder.

• distance.xlsx:
  Geographical distance between any two prefectures.
• gadm36_JPN.gpkg:
  GeoPackage data of the entire Japan, downloaded from Database of Global Administrative Areas (GADM).
• passenger_traffic_2019.xlsx:
  The survey data of passenger traffic between prefectures in 2019, provided by Ministry of Land, Infrastructure, Transport, and Tourism.
• population.csv:
  Population in each prefecture, provided by Statistics Bureau of Japan.
• prefectures.csv:
  Infection status by prefecture, provided by Toyo Keizai Inc.