R Packages on Github

get_r_repos.R

# Get repos with language = R and a given search term, for a given page
get_page <- function(search_term, page){
  require(httr)
  url = sprintf('https://api.github.com/legacy/repos/search/%s?language=R&start_page=%s', 
   search_term, page)
  repos = GET(url)
  return(content(repos)$repositories)
}

# check number of pages directly on github.com, since each page only returns 100 repos in the API call 
RESULTS = 1613 # for package
PERPAGE = 100
pages = seq(1, ceiling(RESULTS/PERPAGE))

# assume package is defined by R repo with package in description
# other terms are NAMESPACE and INST
repos = Reduce('c', lapply(pages, get_page, search_term = 'package'))

# filter out forks to avoid double counting.
repos_orig = Filter(function(x) !(x$fork), repos)

# convert list of lists to a data frame of repos.
repos_data = do.call(rbind, repos_orig)

NOTE. This list is not accurate as it misses out several packages on github. The key is to tweak search_term and come up with something that would account for all packages.

Results of running this code available as a Google Spreadsheet at http://goo.gl/JVZHc