process_data()

process_data() converts a count-based data frame to one row per icon. group_var and sum_var are unquoted (tidy eval); high_group_var takes a character string.

df_raw <- data.frame(
  sex = c("Female", "Male"),
  n   = c(55, 45)
)

df_plot <- process_data(
  data        = df_raw,
  group_var   = sex,
  sum_var     = n,
  sample_size = 20
)

head(df_plot, 4)
#>     type  n prop
#> 1   Male 45 0.45
#> 2   Male 45 0.45
#> 3   Male 45 0.45
#> 4 Female 55 0.55

The result has one row per icon. The type column carries the original group label and is what you map to color and icon in geom_pop().

With high_group_var

Use high_group_var when your data has a faceting variable. It samples independently within each panel so proportions are correct per group.

df_region_raw <- data.frame(
  region = c("North", "North", "South", "South"),
  sex    = c("Female", "Male", "Female", "Male"),
  n      = c(30, 20, 25, 25)
)

df_region <- process_data(
  data           = df_region_raw,
  group_var      = sex,
  sum_var        = n,
  sample_size    = 20,
  high_group_var = "region"
)

head(df_region, 4)
#>   group   type  n prop
#> 1 North Female 30  0.6
#> 2 North Female 30  0.6
#> 3 North   Male 20  0.4
#> 4 North Female 30  0.6

The group column in the output contains the panel label — pass it to facet_wrap(~group) and geom_pop(facet = group).