2  Appendix A: Tidyverse Gallery Scrape

The following code is designed to scrape the names of packages in the ggplot environment from the tidyverse gallery (https://exts.ggplot2.tidyverse.org/gallery/) and then retrieve the cumulative historical cran download count for each. It also scrapes the github star count listed for these packages.

First import the necessary packages:

library(tidyverse)
library(rvest) #html webscraping
library(packageRank) #to retrieve CRAN download counts

Read in the downloaded tidyverse gallery page html file. Must download webpage for the github star scrape to work properly.

df <- read_html("raw_data/exts.ggplot2.tidyverse.org.html")

Scrape the package names and store in a package_names vector.

package_names <- df |>
  html_elements("div.card-content") |> 
  html_elements("span.card-title") |> 
  html_text()

To find the most current total historical download count, set a target_date of two days before today. Depending on the time of day, cranDownloads is updated to either 1 or 2 days previous to the current day.

target_date <- Sys.Date()-2

The below function get_total_downloads takes in a package name to retrieve a cumulative count of that package’s cran downloads up until the set target_date by utilizing the cranDownloads funcion of packageRank. Handles error that arises when package is not found on CRAN. Returned as dataframe.

get_total_downloads <- function(pkg) {
  
  #to = 2025 pulls entire download history
  cd <- tryCatch(
    cranDownloads(packages = pkg, to = 2025),
    
    #if the package is not found in cran return NA
    error = function(e) NA
  )
  
  #retrieving the 'cumulative' value of a particular date gets total download count up      to that date
  count <- ifelse(length(cd) == 1, NA, cd$cranlogs.data$cumulative [ 
    cd$cranlogs.data$date == target_date
  ])
  
  data.frame(package = pkg, downloads = count)
}

Retrieve cran download count for each package by mapping get_total_downloads across scraped package_names and combining returned dataframes in one df. Will take a few minutes to complete.

gallery_packages <- map_dfr(package_names, get_total_downloads)

Scrape github star count from the gallery webpage and add to dataframe.

github_stars <- df |> 
  html_elements("span.github-btn") |>
  html_elements("a.gh-count") |> 
  html_text() |> as.numeric()

gallery_packages$stars = github_stars

Store data in a new dataframe with a column indicating gallery as the source and export as csv with a timestamp for when the data was generated.

gallery_packages$gallery = TRUE

head(gallery_packages)

   package downloads stars gallery
1 ggQQunif     24841     8    TRUE
2  ggupset    182240   354    TRUE
3     xmrr     26242     7    TRUE
4    ggpcp     18302     1    TRUE
5     gg3D        NA   104    TRUE
6     ggQC     63896    46    TRUE

timestamp <- paste("# Gallery data generated on:", Sys.time())

data_lines <- format_csv(gallery_packages)

# Combine and write
write_lines(c(timestamp, data_lines), "generated_data/gallery_packages.csv")