CovMitigation: CovMitigation: Data and modeling for UVAHS epidemic response
In rplzzz/CovMitigation: COVID-19 infection model for Virginia

Modeling Data

The main modeling function is run_scenario. This function takes a vector of time values(in days) to output and a list of parameters. It produces a dataset giving key results, including new infections and acute and ICU admissions. Each call to the function runs a single scenario, which is tagged with a name in the scenario column. Therefore, concatenating the data frames from several runs and plotting with the color aesthetic set to "scenario" is an easy way to compare several different parameter runs.

The second important function is gen_post, which generates a function that takes a vector of model parameters, runs the model and compares to observed data, computes a log-likelihood, and adds a log-prior to get the log-posterior probability density. Embedded in this function is a secondary model for generating the expected number of confirmed cases, given the model outputs. Among other things, this calculation requires us to estimate the effect of targeted testing in raising the number of expected cases, relative to what would be expected from a random sample.

Most of the package data is preprocessed data used in the modeling. This includes information such as demographic data for the local jurisdictions in Virginia, historical UVAHS market shares by local jurisdiction, and so forth.

COVID-19 specific data is hosted in external packages. We use the county-level incidence reports from the New York Times, provided by the NYTimesCOVID19 package, and we use the state level testing counts from the COVID Tracking Project, provided by the vacovdata package. Further information about those datasets can be found in the package documentation for those packages.

rplzzz/CovMitigation documentation built on June 7, 2021, 8:48 a.m.