When you tweet upon a star

I have an ongoing search for tweets containing either the word astronomy or astrophysics, which lets me look at how often the tweet "According to Astronomy, when you wish upon a star, you're actually a few million years too late. That star is dead. Just like your dreams.", and variants, is repeated or retweeted. The graph, wish-upon-a-star-tweet-times.svg, shows the number of all tweets per day that contain either astronomy or astrophysics (red bars) and those that contain the phrase "According to Astronomy, when you wish upon a star" (blue bars). Unfortunately the search was not running for the whole period of the graph, which is why the count drops to near zero at times. The graph is missing a small number of tweets from before the start of July this year (the X axis is labelled year/month/day), just to keep the data readable. Of the 254729 tweets, 15668 are copies of this tweet; that is, 6 percent of Astronomy and Astrophysics tweets! Amazingly enough, more than half the tweets on Saturday, September 14th 2013 were about this. See the referenced articles, by my colleague Karen Masters and Phil 'Bad Astronomer' Plait, for more information on why this tweet is not scientifically accurate, and to see how long this has been floating around the internet. The tab-separated file tweet-upon-a-star-top20.tsv shows those tweets that were replied to the most during the search; this only uses the information provided by Twitter during the search, so can miss some links between tweets depending on how they were created or modified. It shows, amongst other things, that: 1 - many accounts are linked to the spread of this information 2 - the same account (handle in the table) repeats the same information at different times 3 - the third- and fourth-most repeated tweets are for a much-more interesting subject, that of Cecilia Payne-Gaposchkin's contributions to Astrophysics. Please note that, as with any Twitter study, the results here are incomplete since I have no guarantee that the search I am running returns all matching tweets. If the fraction of general astronomy and astrophysics tweets returned is different than that of the "wish upon a star" tweets then this would skew the results.