Analyzing social sentiment

Thanks to Twitter, 140 little characters can make a big impression. Today, those public tweets are pixels in a very big picture. And students at the Annenberg Innovation Lab at the University of Southern California (link resides outside of ibm.com), working with IBM, have turned that picture into meaningful data with a Social Sentiment Index.

The lab was established in 2010 to conduct student-led research that advances new techniques for communications and journalism.

IBM's John Squire highlights the importance of social sentiment and the value that analysis can bring to businesses of all sizes across industries.

The collaboration is aimed at broadening student skills in analytics and demonstrating how Watson-like technologies, such as sophisticated semantic and linguistic analysis software, can crunch Big Data in real-time. Students at define the areas of research and then use IBM's technology to explore how these capabilities can be used by organizations from news outlets and journalists, movie studios and film marketers, retailers and branding experts, in order to better understand, respond and predict public preferences.

The Social Sentiment Index is one such initiative that uses Twitter. With an estimated 155 million tweets per day as of mid-2011, the micro-blogging platform is an excellent tool for sentiment analysis¹.

The Index uses IBM analytics technology to analyze millions of public tweets, with the goal of creating real-time public opinion snapshots. They transform 140 characters from raw, unstructured data into valuable insights, gauging sentiment and following trends on a variety of topics from the retail, sports and entertainment, including major events like the World Series, the Super Bowl and the Academy Awards.

With this real-time analysis, the data churn is fast. A tweet goes out and, micro-seconds later, the sentiment is measured. IBM and USC are even using the new technologies to distinguish irony and apply "machine learning" to figure out which tweets are just background noise—and which truly count.

An eye on the Twitterverse is an ear to the ground

The prevalence of social media means organizations have access to an unfiltered voice ripe for analysis. This requires new marketing approaches, tools and skills to effectively reach customers due to rising online, mobile and social trends. Businesses, academics and journalists can use analytics to gain new insights into consumer perceptions via social media—determining product placement, launch dates and market reception in a quicker, more efficient manner than focus groups or panels.

Professor Jonathan Taplin, director of the Annenberg Innovation Lab, sees potential for students and businesses alike using social sentiment analysis. "The work we are doing with IBM enables our students a unique opportunity to gain valuable knowledge in the use of advanced analytics technologies," he said, "and apply it to real-world settings to understand how this new information can benefit a variety of industries."

In particular, these tools can be used by retailers and sports franchises to better measure and understand consumer sentiment toward products, services and athletes. The idea is similar to a Q rating, or what IBM has dubbed a 'T Score' (Twitter score), that takes brand measurement one step further. By gaining real-time insights into customer and fan preferences news organizations and product marketers can use social media to improve how they serve their respective constituents.

A recent IBM global study of more than 1,700 chief marketing officers revealed that most of the world’s top marketing executives recognize a critical and permanent shift occurring in the way they engage with their customers. And Forrester Research estimates that market opportunity for social software is expected to exceed US$6 billion dollars by 2016, increasing 60 percent each year since 2010. ²

World Series, play by play and tweet by tweet

Goal:
Determine fans’ favorite players as the St. Louis Cardinals played the Texas Rangers in baseball's 2011 World Series.

Result:
The night before the deciding game, the Rangers garnered some 56,600 sentimental tweets, five times the Cardinals' 11,500, and Josh Hamilton was shown to be the fans’ sentimental pick.

Pre-Super Bowl sentiment picks an MVP

Goal:
Determine which quarterback was the fan favorite going into Super Bowl XLVI between the New England Patriots and the New York Giants.

Result:
Overnight results of pre-Super Bowl Twitter buzz drove Giants quarterback Eli Manning’s positive sentiment ahead of Patriot’s quarterback Tom Brady.

And the Oscar® ought to go to...

The Oscar Senti-meter is a tool developed by the L.A. Times, IBM and the University of Southern California Annenberg Innovation Lab to analyze opinions about the Academy Awards race shared in millions of public messages on Twitter.

Goal:
The Senti-meter combs through a high volume of tweets daily and uses language-recognition technology to gauge positive, negative and neutral public opinions shared in the messages. This project illustrates how understanding public social media data can better inform reporting and provide an added dimension to journalist story telling

Result:
Cataloguing these tweets over time gives insight into the vox pop surrounding Hollywood’s awards season and gives a voice to average fans who may or may not endorse the selections made by Tinseltown’s elite. The search terms and key words are being continuously updated so as to capture the most relevant tweets about movies, actors and actresses and their Oscar buzz.

1 Mediabistro, "Twitter Sees 155 Million Tweets Per Day [Stats]", April 4, 2011.

2 PC World, "Forrester: Enterprise Social Software to Become a $6.4 Billion Market in 2016" December 2, 2011.