Twitter sentiments based suggestive framework to predict trends

Abstract

In recent years the increasing computer power and availability of social networking sites such as Twitter, Facebook, Flickr etc. has emerged as an effective medium to express instant views about commercial products, social conditions, political situations etc. Twitter is one such powerful source of information, which can be utilized to investigate the opinion of users. This paper presents an efficient visualization-based framework, which detects opinion from tweets employing lexical analysis approach. The approach is illustrated by taking the example of film industry, where views and reviews affect the opinion about a movie and hence the box office collection. The results obtained were found to be in corroboration with the box office collections. This work is a crucial endeavor to complement the existing techniques of data analysis with the pre-suggestive measures of visualizing the real time opinion of user.