font_df <- hershey %>% 
  filter(font == 'futuram')

head(font_df)
#> # A tibble: 6 x 11
#>       x     y  left right width stroke   idx glyph font    ascii char 
#>   <int> <int> <int> <int> <int> <fct>  <int> <int> <chr>   <int> <chr>
#> 1     0     0    -8     8    16 0          1     1 futuram    32 " "  
#> 2     0    12    -5     6    11 0          1     2 futuram    33 !    
#> 3     0    -2    -5     6    11 0          2     2 futuram    33 !    
#> 4     1    -2    -5     6    11 0          3     2 futuram    33 !    
#> 5     0    12    -5     6    11 1          4     2 futuram    33 !    
#> 6     1    12    -5     6    11 1          5     2 futuram    33 !
ggplot(font_df) + 
  geom_path(aes(x, y, group = stroke)) + 
  coord_equal() + 
  theme_void() + 
  facet_wrap(~glyph, labeller = label_both)
#> geom_path: Each group consists of only one observation. Do you need to
#> adjust the group aesthetic?