packageDownloadStat: Simple Package Download Statistics from CRAN

Description

This function allows accessing the most recent counts of package downloads availabale on http://www.datasciencemeta.com/rpackages, obtaining rank quantiles and to compare (multiple) given packages to the bulk data, optionally a plot can be drawn.

Usage

packageDownloadStat(
  queryPackages = c("wrMisc", "wrProteo", "cif", "bcv", "FinCovRegularization"),
  countUrl = "http://www.datasciencemeta.com/rpackages",
  refQuant = (1:10)/10,
  figure = TRUE,
  log = "",
  silent = FALSE,
  callFrom = NULL,
  debug = FALSE
)

Value

This function retuns a matrix with download counts (or NULL if the web-site can't be accessed or the query-packages are not found there)

Arguments

queryPackages: (character or integer) package names of interest, if integer, n random packages will be picked by random
countUrl: (character) the url where the dayly counts ara available
refQuant: (numeric) add reference quantile values to output matrix
figure: (logical) decide of figure should be printed
log: (character) set count-axis of figure to linear or log-scale (by setting log="y")
silent: (logical) suppress messages
callFrom: (character) allow easier tracking of messages produced
debug: (logical) additional messages for debugging

Details

The task of checking the number of downloads for a given package has been addressed by several packages (eg dlstats, cranlogs, adjustedcranlogs). Detailed articles on this subject have been published on R-Hub (https://blog.r-hub.io/2020/05/11/packagerank-intro/) and on R-bloggers (https://www.r-bloggers.com/2020/10/a-cran-downloads-experiment/).

This function only allows accessing the most recent counts as listed on the website of www.datasciencemeta.com . Please note, that reading all lines from the website may take a few seconds. To get a better understanding of the counts read, reference quantiles for download-counts get added by default (see argument refQuant). The (optional) figure can be drawn in linear scale (default, with minor zoom to lower number of counts) or in log (necessary for proper display of the entire range of counts), by setting the argument log="y".

The number of downloads counted by RStudio may not be a perfect measure for the actual usage/popularity of a given package, the articles cited above discuss this in more detail. For example, multiple downloads from the same IP or subsequent downloads of multiple (older) versions of the same package are counted, too.

Examples

Run this code

## Let's try a microscopic test-file (NOT representative for true up to date counts !!)
pack1 <- c("cif", "bcv", "FinCovRegularization", "wrMisc", "wrProteo")
testFi <- file.path(system.file("extdata", package="wrMisc"), "rpackagesMicro.html")
packageDownloadStat(pack1, countUrl=testFi, log="y", figure=FALSE)
## For real online counting simply drop the argument countUrl