TwitSenti: A Real-Time Twitter Sentiment Analysis and Visualization Framework

Abstract

Twitter is considered as one of the world's largest social networking sites which allow users to customize their public profile, connect with others and interact with connected users. The proposed work introduces a distributed real-time twitter sentiment analysis and visualization framework by implementing novel algorithms for twitter sentiment analysis called Emotion-Polarity-SentiWordNet. The framework is applied to build an interactive web application called "TwitSenti" which can benefit companies and other organizations in knowing the people's sentiment towards the aspects such as brands, current events, etc., which in turn helps in quick decision-making and planning marketing strategies. The algorithm is validated against three existing classifiers and hence proved that Emotion-Polarity-SentiWordNet provides highest accuracy value of 85%. Also, the framework showed best scalability results when evaluated through web app as four node clusters, proves to be fast and can scale well with massive data.