BiologicalRecordsCentre / BRCindicators

An R package for creating indicators from trends data
4 stars 11 forks source link

Species growth rate estimates - option to omit imputed values #79

Open drnickisaac opened 3 years ago

drnickisaac commented 3 years ago

The method produces a species growth rate for each species in each year: these are imputed for species-year combinations with no data. Both "real" and imputed growth rates are included in the spgrowth object, which we use for bar charts in UKBI.

However, for UKBI it does not make sense to include these imputed values for categorising species. I recall that @03rcooke used a workaround for either the Scottish or English indicator set. Could this be easily incorporated into the package?

To avoid drastically changing the behaviour of the method, I would suggest that we add a variable to indicate which growth rates are estimated from data vs imputed, rather than hard coding the removal of imputed values* or insisting that users define what they want a priori.

*it's sometimes of interest to examine the imputed values, e.g. to test whether there may be biases introduced by species with incomplete time-series.

03rcooke commented 3 years ago

Hmm, I don't actually recall this. I don't think I realised that there would be the possibility to separate "real" and imputed growth rates. I thought the spgrowth object just included modelled growth rates, so I treated them all the same. The workaround I built was to accommodate the bma method into the species and indicator assessments, which I have now incorporated into the BRCindicators package.