I wrote a Plotly Reactive Dash app to run the Coronavirus operational hindcast/forecast model written in R code by James D. Annan and Julia C. Hargreaves.
The medRxiv preprint describing their approach is described here:
Model calibration, nowcasting, and operational prediction of the COVID-19 pandemic, James D. Annan and Julia C. Hargreaves (2020)
James and Julia’s hindcast/forecast model is important as it effectively recalibrates the SEIR model as each daily update becomes available.
- Source of population data: World Bank
- Source of daily loss data: CSSE at Johns Hopkins University
- Source of intervention data: Olivier Lejeune visualised using the Global Coronavirus lockdown status app.
The hindcast/forecast model simulation is manifested by 5000 runs using Monte Carlo Markov chains for 6 model parameters:
|Variable 1||Latent period [days]|
|Variable 2||Infectious period [days]|
|Variable 3||0.5 x Initial infection rate|
|Variable 4||Death rate|
|Variable 5||Initial reproductive rate, R0|
|Variable 6||Post-lockdown reproductive rate, Rt|
whose MCMC traces, kernel density estimates and correlation matrix are displayed in the app which is up and running now on Heroku:
and here is the codebase for issue checking on Github: