Nowadays in this digital world we see huge amount of data being created every day, Amazon is one of the leading e-commerce companies which possess such kind of data and Twitter is a famous micro blogging service where its users express their opinion on various topics as “tweets”. We analyze these customer review data to help the customer to come to a conclusion for their purchases. The purpose of this paper is to help users who are trying to buy a new book by providing public opinion based on the Amazon user reviews by constructing an algorithm that can accurately classify sentiments in reviews and also to classify the tweets about those books. Main idea is that we can obtain this high accuracy on classifying sentiments in reviews using natural language processing and machine learning techniques such as bag-of- words, n-gram and Naive Bayes Classifier etc.,. Amazon review data for books for the past decade is itself more than 9GB it’s more than billions on reviews from user around the globe to analyze it and return the most spoken feature about the product we are implementing hadoop technology to make it quick and feasible. This paper may also help the authors, publishers and researchers who want to know the public opinion of the book. The user sentiments will be broadly classified into three categories positive, negative and neutral. Top features of the book (product) will be used to make a user attractive word cloud. © IAEME Publication.