Twitter is growing exponentially, and it’s showing no signs of slowing up. Just last week, it passed the 20-billion tweet mark. For researchers, that means another 2.8 trillion characters of data to analyze.
It feels like almost every day we read yet another report aimed at charting the next big trend on Twitter. Today, for instance, HP released a study claiming to have analyzed enough tweets to discover the most influential Twitter users on the Web. In a realm of 20 billion tweets, how many does it take to draw such conclusions?
The Most Influential Tweeters…HP Labs (August 2010)…22 million tweets analyzed
Business: Avoid Twitter…360i (August 2010)…1,800 tweets analyzed
Twitter’s Oncology Community…JuvoLab (July 2010)…4,450 tweets
The World’s Mood on Twitter…Northwestern University (July 2010)…300 million tweets
Twitter Celebrities are Happy…Edinburgh University (April 2010)…13 Twitter accounts
International Twitter Use…Semiocast (March 2010)…13.5 million tweets
Twitter Users are Inactive…RJMetrics (January 2010)…2 million tweets
40% of Twitter is “Pointless Babble”…Pear Analytics (August 2009)…2,000 tweets
Twitter Users Don’t Tweet…HubSpot (June 2009)…4.5 million Twitter accounts
Twitter Lexicography…Oxford University Press (June 2009)…1.5 million tweets
Few Twitter Users Generate Content…Harvard (June 2009)…300,000 Twitter accounts
The Science of ReTweets…Dan Zarrella (October 2009)…5 million tweets
These somewhat grandiose conclusions are drawn (“analyzed”) from a data sampling anywhere from 13 Twitter accounts to 300 million tweets. Should we believe such conclusions? How much “analysis” can even be done on 140 characters?
With Twitter receiving well over 50 million tweets per day, perhaps the network has grown too big to study.AC