An in-development R package and a Bayesian hierarchical model jointly fitting multiple "local" wastewater data streams and "global" case count data to produce nowcasts and forecasts of both observations
Use the stan spatial functions to implement the generative model of the spatially correlated exponential decay process into the wwinference.stan model, using the non-centered parameterization described here #41.
Requirements
[x] Be able to pass in an adjacency matrix (distance) as data and estimate the correlation function parameters and the spatial deviations in $R(t)$ from simulated admissions and wastewater data
[x] model should compile
[x] get_stan_data() should be modified to pass in distance data
[x] As a first pass, start by assuming we will always have distance data and will always model with a spatially correlated model
Out of scope (for separate PRs)
[ ] modify vignette to describe the distance data (as this represents a new input dataset). Make some figures in the vignette that show the spatial correlation (will have to brainstorm on this as we don't have actual maps, just x and y axis points)
[x] add functionality in pre-processing/stan model to turn the spatial component on or off/
Goal
Use the stan spatial functions to implement the generative model of the spatially correlated exponential decay process into the
wwinference.stanmodel, using the non-centered parameterization described here #41.Requirements
get_stan_data()should be modified to pass in distance dataOut of scope (for separate PRs)