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
Feedback from @ChiragKumar9 indicates it was hard to follow the local vs global variables distinction. Suggests add a diagram showing site level wastewater data and state level hospital admissions + parameters/ delay distributions as model inputs. Consider using a different word other than global variables because intuitively this might imply things like the latent or infectious period, when we are really referring to subsets and total populations within a geographic region.
I think adding a diagram to the vignette (perhaps the schematic we will use for Fig 1 of the manuscript, that itself need revising) would be helpful. Then we want to define clearly what we mean by "global" and "local" variables (or nix that terminology all together).
Goal
Feedback from @ChiragKumar9 indicates it was hard to follow the local vs global variables distinction. Suggests add a diagram showing site level wastewater data and state level hospital admissions + parameters/ delay distributions as model inputs. Consider using a different word other than global variables because intuitively this might imply things like the latent or infectious period, when we are really referring to subsets and total populations within a geographic region.
I think adding a diagram to the vignette (perhaps the schematic we will use for Fig 1 of the manuscript, that itself need revising) would be helpful. Then we want to define clearly what we mean by "global" and "local" variables (or nix that terminology all together).
This probably worth having a discussion on.