Open sbfnk opened 9 months ago
Agree that we could move to a neutral default value and agree we should ideally allow the user to specify this with something else when inputting data. Also agree we should then keep track of this and use where it is needed.
Pushing this to a later version to be addressed alongside #346
At the moment the main functions in the package expect a data frame with columns
date
andconfirm
. Also much of the reporting refers to "confirmed cases". However, the package functionality could be applied to any data streams (e.g. suspect/probable cases, hospitalisations etc.), where "confirmed" cases often refer to a particular subclass that were e.g. lab-tested.I'd suggest to remove the language around "confirmed" cases and perhaps just to refer to "cases", which leaves the case definition open. Alternatively/additionally we could make it a user argument (perhaps as part of #346) so that plot labels, summaries etc. could reflect whatever the model is being fitted to. The input data frame could then also be expected to have a column of more neutral name (e.g.
value
) instead ofconfirm
.