Researchers at Carnegie Mellon University used simple text analysis methods to look at sentiments expressed in one billion Twitter messages between 2008 and 2009, and found that people’s attitudes on consumer confidence and presidential job approval were similar to the results generated by well-reputed, telephone-conducted public opinion polls, such as those conducted by Reuters, Gallup and pollster.com.
The CMU computer science department team filtered updates about the economy and politics and used text analysis to determine if the overall sentiment of the update was positive or negative. The Twitter sentiments had more day-to-day variation compared to data gathered from traditional polling data, so the team averaged the Twitter results over a number of days at which points the results were quite similar to results from polling data.
The paper is available online (use the left column to navigate to the “papers” section, and scroll down to “From Tweets to Polls: Linking Text Sentiment to Public Opinion Time Series”).
Read additional reporting in Mashable’s article.
