Twitter, like many products, is made more relevant and powerful the more users it has and the more active those users become. With over 500 million members sending over 340 million tweets a day, Twitter has become, for many, the go-to destination for news, trends and opinion. While some simply watch as the Twitter machine grows, others choose to mine the information from those hundreds of millions of tweets to gather valuable data that they can use to gauge and predict.
On Tuesday, the Twitter analytics tool Topsy announced a pro version of their product for brand to better analyze search terms, hashtags and trends over time. Topsy is the only company given access to Twitter's firehose, the full stream of data from Twitter, and with that information has been able to record years of tweets for users to mine.
In an article by Business Insider, the CEO of Topsy revealed that one of their customers - a major Hollywood film studio - used Twitter data to decide which actor they should cast for the lead in an upcoming movie. The article states that while for years studios have used the data to predict box office numbers, they've only just recently used the number of @mentions as a deciding factor in their employment.
This year, the company has also teamed up with Twitter to release a political barometer called the Twitter Political Index to try to measure the public's opinion of both presidential candidates. The numbers, set side-by-side, indicate that tweets about the candidates are, on average, more positive than the average Twitter Tweet. In a blog post by the company, they explain, "A score of 73 for a candidate indicates that Tweets containing their name or account name are on average more positive than 73 percent of all Tweets."
While Topsy certainly seems to be at the forefront of Twitter analytics, others have used the tweet info in other creative ways. Researchers at the University of Rochester have built an application that measures health insights in real-time. By analyzing 4.4. million geolocated Tweets in New York City for one month, the group was able to create an algorithm finding tweets specifically about being sick. The data is presented in a heat-map-style data visualization that theoretically can predict sickness.
Many people look at the boundless ocean of Twitter information as a crowded, noisy mob trying to scream over one another, not really getting anywhere but broadcasting opinions to the masses. Luckily not all of us think like that and where one person sees a noisy mob, the next sees a data pool that's every researcher's dream. As Twitter grows and our ability to analyze its data becomes more refined there's really no telling what we may be able to predict from the 340 million tweets that come rolling in everyday.
This is a post from Daniel Levine, a product analyst at Grovo.com, a field guide to the Internet where users can learn about everything from online training to how to use Twitter.