Cannot use weights="area" with weighted_frac

[goergen95]

Hi,

thank you for this amazing work!!

I am trying to obtain a weighted_frac summary with weights = "area". The docs state:

https://github.com/isciences/exactextractr/blob/efa5d50d0301df1477a883ac6f6c504c1cf5d8a1/R/exact_extract.R#L172-L175

and:

https://github.com/isciences/exactextractr/blob/efa5d50d0301df1477a883ac6f6c504c1cf5d8a1/R/exact_extract.R#L134-L136

So, given the repex below, I assume that we currently cannot obtain area weighted summaries for weighted_frac? If that is actually the case, would it be possible to add that feature?

library(exactextractr)
library(sf)
#> Linking to GEOS 3.12.2, GDAL 3.9.1, PROJ 9.4.1; sf_use_s2() is TRUE
library(terra)
#> terra 1.7.78

f <- system.file("ex/elev.tif", package="terra")
r <- rast(f)
s <- scales::pretty_breaks(n = 5)
breaks <- s(values(r))
r2 <- classify(r, breaks)
v <- vect(ext(r2))
v <- st_as_sf(v)
st_crs(v) <- st_crs(r)

exact_extract(r2, v, fun =  "frac")
#>       frac_0    frac_1    frac_2    frac_3     frac_4
#> 1 0.02083333 0.2858073 0.4292535 0.2419705 0.02213542
exact_extract(r2, v, fun =  "weighted_frac", weights = "area")
#>   weighted_frac_0 weighted_frac_1 weighted_frac_2 weighted_frac_3
#> 1      0.02089308       0.2863966       0.4296582       0.2410372
#>   weighted_frac_4
#> 1      0.02201497

^{Created on 2024-08-10 with reprex v2.1.0}

[dbaston]

This appears to be working as expected. There isn't much variation in cell area over this polygon, so taking it into account doesn't change the results by much.

[goergen95]

You are right, I now see that this indeed works as documented. I guess what I am asking is if there is a way to calculate the area per category without using an UDF? Initially, I thought combining weighted_frac with weights = "area" would do this. Sorry for the noise! Please feel free to close this issue anytime.

[dbaston]

Maybe simpler than using a UDF would be to convert the fraction to an area after the fact, like

library(dplyr)
library(stringr)

v$area <- st_area(v)
exact_extract(r2, v, fun='frac', append_cols='area') |>
  mutate(across(starts_with('frac'),
                ~ .x * area)) |>
  rename_with(str_replace, 'frac', 'area', .cols=starts_with('frac')) |>
  select(-area)

#           area_0           area_1           area_2           area_3          area_4
# 1 98683299 [m^2] 1353811417 [m^2] 2033287011 [m^2] 1146165331 [m^2] 104850999 [m^2]

[goergen95]

Yes, this does work, however, requires additional code. Over at mapme.biodiversity we frequently are interested in calculating the area per category and I was hoping this was possible in a more generalized way. Would love to see this implemented here (e.g. something like frac_coverage or similar), but I also understand if you consider this out of scope.

[dbaston]

I've added an upstream issue for this.

[goergen95]

Awesome, thanks!