Analysis of the top R packages

[This article was first published on R on Aurélien Callens, and kindly contributed to R-bloggers]. (You can report issue about the content on this page here)
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.

After a little while coding in Python every day for my work, I needed to make a break and perform some R analysis! Since the beginning of my postdoc, I haven’t followed the last trends concerning R packages. In this post, I am going to analyze some data about R packages to see what are the most downloaded packages during the past weeks. I will also visualize all the relationships between the R packages by looking at their required dependencies.

Let’s import the packages required for this analysis:


Visualisation of the dependencies between packages

Once I saw the above graph, I was wondering about all the dependencies between these packages and I wanted to know which one was the most “connected”. To answer this question, I need more data, especially about the required dependencies of each package. After some research, I found out that data about package (including description and dependencies) can be extracted with a function in the tools package:

df_pkg <- tools::CRAN_package_db()[, c('Package', 'Description', 'Imports')]

However, this function extract the data for all the packages and I want to perform the analysis only on the top 100 popular packages. So I decided to couple the function of the cranlog package with the database I collected with the CRAN_package_db() function:

# Can be quite long hence the parallel map
plan(multisession, workers = 12)
monthly_dl <- future_map(df_pkg$Package, function(x){sum(cran_downloads(x, 'last-month')$count)})
df_pkg$monthly_dl <- unlist(monthly_dl)
# write_csv(df_pkg, 'R_pkg_dl.csv')

We can then filter by number of downloads and keep only the top 100:

df_pkg <- df_pkg %>% 
  distinct(Package, .keep_all= TRUE) %>% 
  arrange(monthly_dl) %>% 
  top_n(100, monthly_dl)

Now, it is time to prepare the data for a graph visualization. To make a graph, we need two tables. The first one must contain all the relationships between nodes (in our case nodes are packages), it has two columns : ‘from’ and ‘to’. The second table contains only one column with the names of the nodes.

import_cleaning <- function(text){
  text <- gsub('\\s*\\([^\\)]+\\)', '', text)
  text <- gsub('\\n', ' ', text)
  text <- gsub(' ', '', text, fixed = TRUE)
  text <- str_split(text, ',')


test <- df_pkg %>% 
  mutate(cleaned_imports = import_cleaning(Imports))

df_target <- function(x,y){
  df <- expand.grid(from=x, to=unlist(y))

for(i in 1:nrow(test)){
  if(i == 1){
    df_res = df_target(test$Package[i], test$cleaned_imports[i])
      df_res = rbind(df_res, df_target(test$Package[i], test$cleaned_imports[i]))

links <- df_res %>% 
  filter(! | (to == ""))

nodes <- tibble(id=as.character(unique(unlist(df_res))))

Once the two matrices are made, we can interactively visualize the graph network with visNetwork package:

visNetwork(nodes, links) %>%
    visIgraphLayout(type = "full") %>%
        shape = "dot",
        color = list(
            background = "#0085AF",
            border = "#013848",
            highlight = "#FF8000"
        scaling = list(min=2,
                       max = 10),
        shadow = list(enabled = TRUE, size = 10)
    ) %>%
        shadow = FALSE,
        color = list(color = "#0085AF", highlight = "#C62F4B")
    ) %>%
    visOptions(highlightNearest = list(enabled = T, degree = 1, hover = T)) %>% 
    visLayout(randomSeed = 11)

Do not hesitate to move, zoom in or select packages to see their dependencies !

To leave a comment for the author, please follow the link and comment on their blog: R on Aurélien Callens. offers daily e-mail updates about R news and tutorials about learning R and many other topics. Click here if you're looking to post or find an R/data-science job.
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.

Never miss an update!
Subscribe to R-bloggers to receive
e-mails with the latest R posts.
(You will not see this message again.)

Click here to close (This popup will not appear again)