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
The red line is the data, and the other coloured lines I think are draws from the posterior predictive distribution. I think these draws from the posterior predictive distribution should be piecewise constant (like the data). This might be possible using e.g. geom_step (or otherwise not using geom_line to connect up the shortest path).
As I understand it the code for these plots was from somewhere in this package.
62 fixes this. Also noting that it is also important to make sure one is not comparing an infection-feedback-adjusted $\mathcal{R}(t)$, which will indeed vary day-to-day to an unadjusted weekly $\mathcal{R}(t)$
dylanhmorris
commented
2 months ago
In https://github.com/cdcent/cfa-spatial-ww-internal/issues/5#issue-2458143691 @cbernalz created these plots.
The red line is the data, and the other coloured lines I think are draws from the posterior predictive distribution. I think these draws from the posterior predictive distribution should be piecewise constant (like the data). This might be possible using e.g.
geom_step(or otherwise not usinggeom_lineto connect up the shortest path).As I understand it the code for these plots was from somewhere in this package.