We have just merged a new vignette (Estimating the effective reproduction number in real-time for a single timeseries with reporting delays • epinowcast) that discusses how to use
epinowcast to estimate the effective reproduction number whilst adjusting for right truncation (i.e. nowcasting). This is an extension of some earlier work @sbfnk and I did last year (Nowcasting in a real-time analysis pipeline to estimate the effective reproduction number with missing data and reporting delay) but uses the new functionality in
epinowcast to do everything jointly. Opening this thread for any feedback and/or discussion that isn’t suited for an issue (at least not straight away).
This is also very related to a recent preprint from @adrianlison which includes the first (that I know of) model that can jointly estimate R_t and correct for right truncation.
There are a fair few pain points and in many ways this vignette mostly serves to highlight where more work is needed.
There are already a few issues open to attack some of these which others are very welcome to get involved with:
- Check that retrospective nowcasting works as expected · Issue #318 · epinowcast/epinowcast · GitHub
- Add effective reproduction number vignette plots to package as functions · Issue #317 · epinowcast/epinowcast · GitHub
- Porting distribution wrappers into their own package · Issue #316 · epinowcast/epinowcast · GitHub
- Extract work on simulated double censoring into its own lightweight package · Issue #315 · epinowcast/epinowcast · GitHub