How to order facets based on facet_wrap from Y values?

Question

is there a way to reorder those facets based on their Y value (aka count ) in decreasing order dynamically - without hardcoding the order (given a similar column to package with hundreds of levels?

  library(cranlogs)
  library(tidyverse)   

  tidyverse_downloads <- cran_downloads(
      packages = pkgs,
      from     = "2017-01-01",
      to       = "2017-06-30") %>%
      tibble::as_tibble() %>%
      group_by(package)



  tidyverse_downloads %>%
     ggplot(aes(x = date, y = count)) +
     #geom_hline(yintercept = 0, color = "grey40") +
     geom_line() +
     #geom_smooth(method = "loess") +
     theme_minimal() +
     facet_wrap(~ package, scale = "free_y",ncol = 2) + theme(axis.text.x = element_text(angle = 90, hjust = 1))

What I have tried was to use the package library(tidytext) and scale_y_reordered

   tidyverse_downloads %>%
     mutate(count = tidytext::reorder_within(count, date, package)) %>% 
     tidytext::scale_y_reordered() +
     ggplot(aes(x = date, y = count)) +
     geom_line() +
     theme_minimal() +
     facet_wrap(~ package, scale = "free_y",ncol = 2) + theme(axis.text.x = element_text(angle = 90, hjust = 1))

however, that doesn't produce any outcome. Im doing something wrong I guess. Appreciate any help

Answer 1

You can sum the total downloads of each package to gather the factor levels in decreasing order.

library(cranlogs)
library(tidyverse) 

pkgs <- c('dplyr', 'broom', 'ggplot2', 'purrr')
  
tidyverse_downloads <- cran_downloads(
  packages = pkgs,
  from     = "2017-01-01",
  to       = "2017-06-30") %>%
  tibble::as_tibble()


tidyverse_downloads %>%
  group_by(package) %>%
  summarise(count = sum(count)) %>%
  arrange(desc(count)) %>%
  pull(package) -> levels

levels
#[1] "ggplot2" "dplyr"   "purrr"   "broom"  

Assign the levels in package according to levels above.

tidyverse_downloads %>%
  mutate(package = factor(package, levels)) %>%
  ggplot(aes(x = date, y = count)) +
  geom_line() +
  theme_minimal() +
  facet_wrap(~ package, scale = "free_y",ncol = 2) + 
  theme(axis.text.x = element_text(angle = 90, hjust = 1))