Analysis of janitor downloads

janitor_usage_analysis.R

# Exploring download counts of a single package
x <- cranlogs::cran_downloads("janitor", from = "2016-10-03", to = "2021-03-12")

library(tidyverse)
library(lubridate)
library(tntpr) # from devtools::install_github("tntp/tntpr")

x$wday <- wday(x$date)
x$weekday <- ifelse(x$wday %in% c(1,7), "Weekend", "Weekday")
x$year <- tntpr::date_to_sy(x$date, as.Date("2016-10-02")) # segments into years using a cutoff date

ggplot(x, aes(x = date, y = count, color = weekday)) +
  geom_smooth() +
  geom_point() +
  geom_vline(xintercept = c(as.Date("2016-10-03"),
                            as.Date("2016-10-31"),
                            as.Date("2017-05-06"),
                            as.Date("2018-01-04"),
                            as.Date("2018-03-22"),
                            as.Date("2018-07-18"),
                            as.Date("2018-07-31"),
                            as.Date("2019-04-20"),
                            as.Date("2020-01-22"),
                            as.Date("2020-04-08"),
                            as.Date("2020-04-12"),
                            as.Date("2021-01-05")),
             alpha = 0.4) +
  theme_minimal() +
  labs(x = "Date", y = "Daily Downloads", color = "")

x %>%
  group_by(year) %>%
  summarise(dls = sum(count)) %>%
  mutate(growth = dls / lag(dls))

# Rank and percentile vs. all CRAN packages
# Code from http://josiahparry.com/post/cran-2019/

# To download anew - this may run for over an hour
# Or you can download the .csv of this time period from my Github gist, see link below

library(rvest)
library(tidyverse)
url <- "https://cran.r-project.org/web/packages/available_packages_by_name.html"
cran_packages <- html_session(url) %>% 
  html_nodes("table a") %>% 
  html_text()

get_year_downloads <- function(pkg) {
  cranlogs::cran_downloads(pkg, 
                           from = "2021-01-01",
                           to = "2021-03-01") %>% 
    group_by(package) %>% 
    summarise(total_downloads = sum(count))
}
total_cran_downloads_2020 <- furrr::future_map_dfr(
  cran_packages, 
  .f = possibly(get_year_downloads, otherwise = tibble())
)

# For the time period Jan-Feb 2021, I've posted the .csv here:
# https://gist.github.com/sfirke/7654da28c08db62fe70001eac4ecf0fe
# if already downloaded:

total_cran_downloads_2020 <- read_csv("cran_downloads_jan_feb_2021.csv")
total_cran_downloads_2020$percentile <- ecdf(total_cran_downloads_2020$total_downloads)(total_cran_downloads_2020$total_downloads)
total_cran_downloads_2020$rank <- min_rank(desc(total_cran_downloads_2020$total_downloads))
total_cran_downloads_2020 %>%
  filter(package == "janitor")