Particle Markov-chain Monte Carlo with Demographic parameters and the JHU CSSE infections/deaths time series

Hey Everyone,
I am trying to implement a Particle Markov-chain Monte Carlo (PMCMC) model using the SimBIID package in R (https://cran.r-project.org/web/packages/SimBIID/SimBIID.pdf) with the goal of forecasting new infections and deaths for each US county.

I have acquired and combined (based on the column FIPS) the US county-level demographics data from the recent publication “A County-level Dataset for Informing the United States’ Response to COVID-19” from the github here: https://github.com/JieYingWu/COVID-19_US_County-level_Summaries/blob/master/data/counties.csv with the JHU CSSE time series data per US county (https://github.com/CSSEGISandData/COVID-19/tree/master/csse_covid_19_data/csse_covid_19_time_series, confirmed_US and deaths_US).

What I’m having a hard time with is figuring out how to add the demographics data (population density, public transport usage, etc) into the PMCMC model in SimBIID.
Any other general advice on PMCMC for predicting based on a time series of data would be greatly appreciated.

2 Likes