Abstract
As utilization of internet business is expanding at a large domain consumers not only buy products but also give their input recommendations that will be very advantageous to different clients. Individual’s opinion and experience are extremely important information for decision making process. Now a days many websites encourages clients to express and trade their ideas, opinions, views, suggestions related to particular product, its policies, services publically. Sentiment analysis is one of the most sought after technique to review information posted by user to gain insights for decision making process. Data collected for sentiment analysis from heterogeneous sources often comprises of missing values, noisy data etc. which needs to be preprocessed using data dictionary. In this paper we are aiming to provide enhanced data dictionary to handle data preprocessing more efficiently and accurately required for sentiment analysis. We worked on a case study based on twitter data to find the brand reputation of three popular mobile brands based on our enhanced data dictionary.