Ranking Online Consumer Reviews

Abstract

The main Aim of our project is to rank products and services bases on customer and user reviews. For this purpose we already have a google reviews system, but our web application will be different from google reviews. Google review system only focuses on star ratings, thier system not include comments in overall ratings of product or service. On the other hand our web application will consider both star ratings and comments also. Because we have seen many contradicted reviews on different brands and products and we believe that thier star ratings can improve further. Our system will identify contradicted reviews and will separate contradicted reviews from fair reviews, and then we will calculate overall ratings from fair reviews only. To build such type of web application we will use the python programming language. We will use selenium to scrapping data and then we will use NLTK to identify the contradicted reviews and separate them from fair reviews. For front-end purposes, we will use html and css to achive the front end design. We will use two text files, one will contain positive words and the second one will contain negative words, we get those two files from google, they are consist of 100 and thousand positive and negative words. Each comment will be analyzed using these two files to check whether the comment contains a postive comment or a negative word after analyzing comments. This web application will help online consumers to take decision before buying any product or service online. We will add more accuracy to actual ratings of local services or products.