Open andybeet opened 1 month ago
Unsure, but this might be irrelevant. Here's why:
Because the assign_unknowns
is the last step in the process we have already aggregated areas and gears. So we are filling in missing values at the aggreate level, based on aggregated data. We should be doing the "data filling" step BEFORE the aggregation step
In
aggregate_area
, the landings and the value can be proportioned from a statistical area (those overlapping multiple EPUS) among EPUS.This step is completed prior to the
assign_unknown
step which uses data from this previous step to "fill" in missing variables like AREA, MONTH, NEGEAR (or EPU, MONTH, FLEET) ifaggArea
oraggGear
are turned on.In this
assign_unknow
step, if a missing AREA is assigned an AREA value that overlaps two EPUs, are the value and the landings proportioned or is the original landings and value retained? If the latter then we'd be inflating landings and value