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
We want to improve the current method of going from a spatial to non-spatial model in the wwinference stan model.
Context
Currently the way this is implemented, if the user passes in a distance matrix, based of wastewater site locations, the wwinference model will use the spatial model. If the user does not pass in a distance matrix, the wwinference model will use the non-spatial model. We want a switch that the user can pick which spatial / non-spatial model is used regardless of whether or not the user provides a distance matrix.
Goal
We want to improve the current method of going from a spatial to non-spatial model in the wwinference stan model.
Context
Currently the way this is implemented, if the user passes in a distance matrix, based of wastewater site locations, the wwinference model will use the spatial model. If the user does not pass in a distance matrix, the wwinference model will use the non-spatial model. We want a switch that the user can pick which spatial / non-spatial model is used regardless of whether or not the user provides a distance matrix.
Requirements