Sentiment analysis of Amazon product reviews based on NLP

This is the age of information explosion. Numbers, text, video, and audio are all information. How to extract the instructive information we are interested in from a large amount of data is crucial. In this article, we built an evaluative model based on amazon product reviews. The review content is divided into good, medium, and bad to express the positive, negative and neutral emotions. From the basic information about the data, we find that the number of reviews is increasing exponentially. More and more consumers choose to shop online, among which beauty and baby products are more popular among consumers. To understand consumers' preferences, we use TextBlob, a natural language processing library in Python, to do word frequency statistics and sentiment analysis on product reviews. It is found from the word cloud that consumers are most concerned about the quality and price of products. The specific quality description of text-based Reviews - "love" is closely related to the rating level.