About a month ago I was at Great Lakes THAT Camp 2012 (staying in a former convent) and I made a graph of tweeting action and posted about it at #thatcamp hashtags the Tony Hirst way.
The last couple of days I was at Code4Lib North 2012 (staying in a former convent) and on the train back home I made a graph of tweeting action, this time using ggplot2
in R
the way Tony Hirst had. This should let you recreate it:
> require(ggplot2) > require(twitteR) > require(plyr) > tweets <- searchTwitter("#c4ln", n=500) > tw <- twListToDF(tweets) # turn results into a data frame > tw1 <- ddply(tw, .var = "screenName", .fun = function(x) {return(subset(x, created %in% min(created),select=c(screenName,created)))}) > tw2 <- arrange(tw1, -desc(created)) > tw$screenName <- factor(tw$screenName, levels = tw2$screenName) > library(stringr) > trim <- function (x) sub('@','',x) > tw$rt=sapply(tw$text,function(tweet) trim(str_match(tweet,"^RT (@[[:alnum:]_]*)")[2])) > tw$rtt=sapply(tw$rt,function(rt) if (is.na(rt)) 'T' else 'RT') > png(filename = "20120525-c4ln-tweets.png", height=800, width=600) > ggplot(tw)+geom_point(aes(x=created,y=screenName,col=rtt)) > dev.off()
(If you don't have the twitteR
(or ggplot2
or plyr
) package installed then run install.packages("twitteR")
(or the other package name) to get it.)
What does this show us? Original tweets and old-style RT retweets, first off. That a few people started using the hash tag early(@ajmcalorum was first off the bat). That a few people used it a lot (the usernames, lower down, with a lot of dots along their horizontal line) and more people used it only once (usually retweets). That there were a number of tweets on 23 May (the day before, when people were travelling to Windsor to get there), but things started in earnest on the first day, 24 May.