UCLA

With the increasing level of access to online political discourses, made possible

by the social media networks, a systematic analysis of political speech becomes

more critical. The available data could be useful in analyzing important

controversies and estimating the general public's opinion in important political events

. A deeper insight enables us to track the public's opinions to detect the shift of

sentiment and discuss on potential reasons why the shift has happened.

Twitter has thus far, shown to be the most effective social network for political

analysis. Tweets are limited to 140 characters which lead the users to make

succinct statements. The simplicity of text along with the relatively high level of

accessibility has made Twitter, a popular target for social analysts.

In this study, we have provided a mechanism for online and automatic collection

of Twitter data surrounding different people and events. The data is maintained

in an organized time based fashion which makes it easy for future references. We

have applied sentiment analysis to estimate political popularity and track the sentiment over time. Sentiment scores have been defined for multiple politicians in

order to allow for a comparison of popularity.