2023-05-23_r-packages-github.md

Comment

In response to my blog post:

Great analysis! Nicely done! It would be nice to compare the number of commits with the released version of the packages, some packages do a lot of commits for few releases.

https://fosstodon.org/@Lluis_Revilla/110416919345900112

Code

I don't know if there's a faster way to get the number of releases for a package, but here's one approach. It takes around 20-30 minutes.

library(tidyverse)
library(scales)
library(pkgsearch)

# Get CRAN package GitHub contributors data from my post:
# https://tidytales.ca/posts/2023-05-10_r-developers-github/
contributors_data <- "https://raw.githubusercontent.com/mccarthy-m-g/tidytales/main/data/2023-05-10_r-developers-github.csv"
contributors_all <- read_csv(contributors_data)

packages_all <- set_names(unique(contributors_all$package))

# {pkgsearch} can get package histories as a data frame where each row is a new
# CRAN release, so we just need the number of rows for a packages history to
# get the total releases.
package_releases <- map_dfr(
  packages_all,
  \(.x) {
    tryCatch(
      tibble(releases = nrow(cran_package_history(.x))),
      error = function(e) tibble(releases = NA)
    )
  },
  .id = "package"
)

# Get summary statistics for each package
packages_tidy <- contributors_all |>
  group_by(package) |>
  summarise(
    contributors = n(),
    commits = sum(n_commits, na.rm = TRUE)
  ) |>
  left_join(package_releases)

ggplot(packages_tidy, aes(x = releases, y = contributors)) +
  geom_hex(bins = 30) +
  geom_rug(
    alpha = 0.01, position = position_jitter(width = .1, height = .1)
  ) +
  scale_x_continuous(
    trans = "log2", breaks = breaks_log(n = 7, base = 2)
  ) +
  scale_y_continuous(
    trans = "log2", breaks = breaks_log(n = 7, base = 2)
  ) +
  scale_fill_viridis_c(trans = "log10") +
  theme(panel.grid.minor = element_blank())

ggplot(packages_tidy, aes(x = releases, y = commits)) +
  geom_hex(bins = 30) +
  geom_rug(
    alpha = 0.01, position = position_jitter(width = .1, height = .1)
  ) +
  scale_x_continuous(
    trans = "log2", breaks = breaks_log(n = 7, base = 2)
  ) +
  scale_y_continuous(
    trans = "log10", breaks = breaks_log(n = 7, base = 10)
  ) +
  scale_fill_viridis_c(trans = "log2") +
  theme(panel.grid.minor = element_blank())

^{Created on 2023-05-23 with reprex v2.0.2}