I have been a fan of the Journal of Open Source Software for a while. Recently I had the opportunity to review a submission for them, which was a great experience. As of today they have accepted 378 papers! Sometimes I browse the new papers and recently I have noticed that the papers seem to be getting longer. I thought I would investigate this observation by looking at the data:

library(dplyr)
library(purrr)
library(fs)
library(tsibble)
library(pdftools)
library(git2r)

jp <- path_home("Downloads", "joss-papers")

if (!dir_exists(jp)) {
  dir_create(jp)
  clone("https://github.com/openjournals/joss-papers.git", 
      local_path = jp)  
}

paper_length <- jp %>% 
  dir_ls(recursive = TRUE, glob = "*.pdf") %>% 
  map_dbl(~ pdf_info(.x)$pages)

avg_length <- paper_length %>% 
  slide_dbl(mean, .size = 25)

plot(paper_length, pch = 1, cex = 0.3, bty = "L",
     xlab = "Paper Number", ylab = "Page Length")
lines(avg_length, lwd = 3, col = "blue")