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
Currently the model expects daily count data, and we have set it up to throw an error if this is not the case. However, we should be able to let the user specify that the count data is not daily, and then sum up the daily latent hospital admissions to get aggregated to whatever interval (most likely weekly, but this could be configurable).
Tasks
[ ] modify wwinference to let the user specify the temporal granularity of the data
[ ] modify wwinference.stan to aggregate between non successive hosp_times if the user specifies an interval for aggregation (otherwise, a missing hosp time is treated as unobserved)
[ ] write tests for both cases, consider including an example in vignette (though this might be a separate vignette where we demonstrate advanced options?)
Goal
Currently the model expects daily count data, and we have set it up to throw an error if this is not the case. However, we should be able to let the user specify that the count data is not daily, and then sum up the daily latent hospital admissions to get aggregated to whatever interval (most likely weekly, but this could be configurable).
Tasks
wwinferenceto let the user specify the temporal granularity of the datawwinference.stanto aggregate between non successivehosp_timesif the user specifies an interval for aggregation (otherwise, a missing hosp time is treated as unobserved)