Currently, the multi-category stat and geom layers operate in tandem, while the time series stat and geom operate wholly separately. What makes more sense is to have two stat layers, one for each type of data, and only one geom_alluvium(), which is fed an appropriately transformed data frame from either stat layer.
(Still unfamiliar with closing issues.) This was done, a bit differently than described here, in the recent merge from the "lodes" branch. See the new vignette for details.
Currently, the multi-category stat and geom layers operate in tandem, while the time series stat and geom operate wholly separately. What makes more sense is to have two stat layers, one for each type of data, and only one
geom_alluvium()
, which is fed an appropriately transformed data frame from either stat layer.