navel-gazer.R

navel.gazer <- function(month = NULL, year = NULL, entire = FALSE,
                         list = "r-help", n = 50, plot = TRUE) {
  # Ben Bolker came up with most of the code
  # Henrique Dallazuanna provided an edit to the z <- line of code
  # Brian Diggs provided capwords() to properly count Peter Dalgaard
  # Joshua Wiley adapted all of it to one function


  if(is.null(month)) {
    month <- format(Sys.Date(), format = "%B")
  }
  if(isTRUE(entire)) {
    month <- unique(months(as.Date(1:365, "2000-01-01")))
  }
  if(is.null(year)) {
    year <- format(Sys.Date(), format = "%Y")
  }
  if(length(year) > 1) {
    tmp <- vector(mode = "list", length = length(year))
    for(i in seq_along(year)) {
      tmp[[i]] <- paste(year[i], month, sep = "-")
    }
    times <- unlist(tmp)
  } else {
    times <- paste(year, month, sep = "-")
  }

  require(zoo)
  
  times <- sort(as.yearmon(times, "%Y-%B"))
  current <- as.yearmon(Sys.Date(), "%Y-%m")
  times <- format(times[times <= current], "%Y-%B")
  
  # Function to extract the names
  # Originally by Ben Bolker
  namefun <- function(x) {
    gsub("\\n","",gsub("^.+<I>","",gsub("</I>.+$","",x)))
  }

  # Based on a suggestion by Brian Diggs
  # Capitalizes the first letter of each word
  capwords <- function(s, strict = FALSE) {
    cap <- function(s) paste(toupper(substring(s,1,1)),
                         {s <- substring(s,2); if(strict) tolower(s) else s},
                         sep = "", collapse = " " )
    sapply(strsplit(s, split = " "), cap, USE.NAMES = !is.null(names(s)))
  }

  # Collects the author names for the relevant month and list
  # from the R archives
  # Originally by Ben Bolker
  grabber <- function(month, list, n) {
    baseurl <- "https://stat.ethz.ch/pipermail/"
    require(RCurl)
    # z <- getURL(paste(baseurl,list,"/",month,"/author.html",sep=""))
    z <- getURL(paste(baseurl,list,"/", month,"/author.html",sep=""),
                ssl.verifypeer = FALSE)
    zz <- strsplit(z,"<LI>")[[1]]
    cnames <- capwords(sapply(zz[3:(length(zz)-1)],namefun))
    rr <- rev(sort(table(cnames)))
    output <- rr[1:n]
    return(output)
  }

  # Create dot plots of the number of posts
  # lattice dotplot() code primarily by Ben Bolker
  plotter <- function(dat) {
    require(lattice)
    if(length(dat) > 1) {
      old.par <- par(no.readonly = TRUE)
      on.exit(par(old.par))
      par("ask" = TRUE)
    }
    for(i in seq_along(dat)) {
      print(dotplot(~rev(dat[[i]]), xlab = "Number of posts",
                    main = names(dat)[i]))
    }
    invisible()
  }

  numbers <- lapply(times, function(x)
                    {grabber(month = x, list = list, n = n)})
  names(numbers) <- times

  if(plot) {
    plotter(dat = numbers)
  }
  
  return(numbers)
}

Back in August 2010 on the R-help list, Ben Bolker posted some code to look at a graph of the top 50 posters on the main R-help list for a given month (original thread: http://www.mail-archive.com/r-help@r-project.org/msg106571.html ). I have combined his original code along with a few other recommendations from that thread into one easy function. This function has a number of convenient defaults. If no arguments are passed, it will look up the top 50 authors on the r-help list, for the given month in the given year. You can also specify one or more months as a character vector (e.g., "August" or c("August", "September") ). The same goes for years. You can also choose a different list (such as r-devel).

For those with a lot of time one their hands, you can set argument entire = TRUE, which will look up every month from whatever years you specified (or the current year if you did not). It will return a named list with each element corresponding to one month. Also, by default it will create a dotplot in lattice (though this may be turned off via plot = FALSE). Finally, you can specify how many authors you want. It defaults to 50.