Code for the paper "Using mobility data in the design of optimal lockdown strategies for the COVID-19 pandemic"

We report here code, data and results for England; inference is done with data until 23rd May and 31st August.

The preprint which focuses on England can be found at:

https://arxiv.org/pdf/2006.16059.pdf

Please also check the companion website, where updates on subsequent research will be described:

https://optimallockdown.github.io/Covid19inEngland/

Content

The content of this repository is as follows:

• data contains the data used in the paper for model fit and optimal control. It contains both the raw data and the one formatted for using our code.
• plot_routines contains some scripts to produce plots in the paper.
• results contains the .jrl files used to store the inference results in the ABCpy library, together with some figures.
• src contains source code, with model definition and various utilities functions.
• Datasets.ipynb is a jupyter notebook exemplifying the data sources and the data processing operations before fitting the model, for inference with observed data until 23rd May.
• Dataset_England.py is instead a Python script used to format data for inference with observations until 31st August.
• inference_SEI4RD_england_data.py is the script performing model fit; can be used with MPI parallelization by uncommenting one line.
• optimal_control_posterior_mean.py is the script performing optimal control under uncertainty, together with the definition of several functions to perform optimal control in specific cases.

Requirements

The following should install the required packages:

pip3 install -r requirements.txt

According to your python3 installation, you may require pip instead of pip3.