library (twitteR)
library(lattice)

# first we will rewrite Jeff Gentry's Rtweets function to make it more flexible
gettweets <- function (session = getCurlHandle(), tag = "rstats", n = 20, page = 1) 
{
    uri <- paste("http://search.twitter.com/search.json?q=&tag=", tag, "&rpp=", n , "&page=", page, sep="")
    out <- getURL(uri,  curl = session)
    jsonList <- twFromJSON(out)[[1]]
    sapply(jsonList, buildStatus)
}

# now get a bit of data
rtw <- gettweets(n=100, page=1)
rtw <- append(rtw, gettweets(n=100, page=2))

# plot the usernames
tweeters <- sapply(rtw, function (x) slot(x, "screenName"))
trellis.par.set(list(fontsize=list(text=6)))
barchart(sort(table(tweeters)), horiz=T, col="light blue")

# how many users?
length(table(tweeters))

# quantile break to fine users responsible for 50% of the tweets
quantile(tweeters,probs = c(0.5))
panel.abline(h=73, col="red", lty=2) # (you may need to change your parameter for h)


## look for tweets containing links
tweets <- sapply(rtw, function (x) slot(x, "text"))
pct <- round(length(tweets[grep("http",tweets)])/length(tweets)*100, digits=2)

barplot(matrix(c(length(tweets[grep("http",tweets)]),length(tweets)-length(tweets[grep("http",tweets)]))), col=c("brown","beige"))
text(0.7,70, paste ("with links:",pct,"%")) # (you may need to adjust text position)
text(0.7,200, paste ("without links:", 100-pct,"%"))


## look for other hashtags in posts
tokens <- unlist(strsplit(tweets, " "))
hashtags <- tokens[grep("#", tokens)]
barchart(sort(table(hashtags)), horiz=T, col="lightgreen")