WKSP) Sentiment Analysis of Twitter Samples That Differentiates Impact of User Participation Levels

The microblogging social media platform Twitter, accounting for millions of 'tweets' per day, provides an effective platform for sampling conversations on a wide array of topics and influences a variety of research areas. Coupled with the presupposition that online conversations often mirror offline conversations, many researchers leverage Twitter samples to justify conclusions about the larger population. More recently, researchers are sampling Twitter for sentiment analysis or opinion mining on products and services and, relevant to this work, for political and social commentary that may lead to election prediction. Traditionally, sentiment analysis has been visualized as an aggregation of opinions expressed in the content discussed online, while neglecting the presence of the creators of the content and the impact of their varying levels of participation. This paper illustrates and proposes an alternative model for evaluating and visualizing sentiment using Twitter samples while leveraging and highlighting user participation and impact.