In this repository we present Julia notebooks for modeling the COVID-19 breakout in South Africa and outher countries using case counts from open data sets.
This directory contains a selection of Jupyter notebooks with Julia code for estimating Rt in real time. The code uses maximum likelyhood to estimate the "most likely" R value for a population at time t. Version 1 of the code is a direct translation from Kevin Systrom's Python code discribed at http://rt.live/ and made available to the public as a Jupyter notebook with accompanying explanations of the methods employed. Version 2 and 3 are modifications of that code with version 3 implementing a simple fast discrete estimation of Rt from daily case counts.
Below is a visualisation of Rt over time for the COVID-19 South African case counts.
This graphic can be generated using the RtLiveZA.ipynb notebook from the RtLive
directory.
This directory contains a Jupyter notebook that uses julia to perform discrete simulation
and optimisation to fit a simple piecewise SIR model to John Hopkins case count data.
The piecewise nature of the model allows the user to observe the effect of government
interventions. A spreadsheet is also provided to create simple SIR simulations.
Below is a visualisation of the early breakout of COVID-19 in South Africa.
This graphic can be generated using the covid_sir.ipynb notebook from the SIRjulia
directory.