Twitter | Twitter And Our Feelings: Real-Time Sentiment Analysis

Twitter was founded in 2006. Over the course of the last four years, in particular, we've seen a gigantic uptake of the platform by journalists, business leaders, brands, and celebrities. Twitter is a call for help, a way to report news, a way to shout out things happening around you. People also watch events together on television, while tweeting out their quips and thoughts.

Due to the uptake of over 100 million active users on Twitter today and the fact that people are watching events together on the platform (elections, the Oscars, sports events) and making commentary, Francois Bar, Associate Professor of Communication at the University of Southern California and his team of researchers have explored sentiment analysis. The majority of Bar's research is grounded in exploring social and economic impacts of information technologies.

The interaction of millions of individuals through social networks generates vast amounts of data. With that data, teams from the Innovation Lab and at the Viterbi School of Engineering at USC are trying to analyze collective knowledge.

The value here is in determining what forms public opinion and why viewers might feel positively or negatively about a certain candidate, team, or event.Gauging public opinion is not an easy target, but Twitter and the people using it give a real-time view into the collective mind.

From USC's project mission:

"Most Twitter sentiment analysis to date has been post-hoc, providing assessment of political sentiment days or weeks after an event, and often based on a sample of tweets. By contrast, we extract all relevant tweets in real-time."

"Our sentiment analysis model aims to overcome the limitations of other models, which adopt simple binary classification of words or phrases. Our approach analyzes the correlational weight of each word with each emotion category - including positive, negative, neutral, and unsure classes - to allow better modeling of objective tweets and tweets that are difficult to classify."

Watch below, as Bar explains more about sentiment analysis: