Extracting Sentiments from Product Reviews

Deciphering the emotional tone behind a sequence of words finds extensive utility in analyzing survey responses, customer feedback, or product reviews. In particular, the advent of social networks offered new possibilities for people to instantly express their opinions on various issues. Therefore, it is not surprising that many shareholders—such as companies, academia, or government—aim to exploit public opinion on various topics and acquire valuable insight.

This chapter focuses on another typical problem in natural language processing (NLP): the extraction of sentiment from a piece of text. For this reason, we incorporate an open source dataset with customer reviews from the Amazon online store. Exploratory Data Analysis (EDA) is again the first task in the pipeline, which helps us discuss important findings on the input data. During this phase, we create different visualizations and enhance our plot construction skills...