The impact of readers' sentiments on book sales: an applied study of sentiment analysis of bestselling book reviews on Amazon
Year: 2024
Type: article
Abstract: The study aimed to analyze the sentiments of reviews for the bestselling books on Amazon to explore the positive, negative, and neutral emotions expressed in these reviews and assess their impact on book sales. This was achieved through the use of machine learning and natural language processing techniques available in the Orange Data Mining environment, supported by Python libraries and algorithms such as Vader and LDA. The study followed a descriptive-analytical approach and collected a total of 4,145 reviews using web scraping techniques from Amazon. These reviews pertained to the top 20 bestselling books on Amazon during the period from August 18 to August 25, 2024. The study yielded several key findings, the most notable being that the book Good Energy ranked first among the top 20 bestselling books on Amazon, followed by Imminent in second place and Atomic Habits in third. The study also demonstrated that bestselling books are not limited to a specific genre but span various fields of knowledge. One of the main conclusions was that readers' emotions towards books are a significant factor influencing purchasing decisions and book sales. The study revealed that positive sentiments dominated the reviews, with approximately 3,603 positive reviews, representing 87% of the total. Most of these reviews were classified under the "Joy" category, accounting for 3,474 reviews or 83.3%, according to Ekman’s model. The study concluded with a set of recommendations and proposed mechanisms directed at authors, publishers, and digital platforms. One key recommendation is to leverage sentiment analysis results in marketing campaigns and develop recommendation algorithms that do not solely rely on star ratings but also incorporate sentiment analysis of reviews. It also suggested displaying books with higher positive sentiment at the top of the browsing lists in specific subject areas. Additionally, the study recommended issuing periodic reports listing books that recorded the highest percentage of positive sentiment in reader reviews, categorized by subject. It also proposed providing a brief graphical summary for each book, showing the evolution of reader sentiments and how they reflect on the book’s sales performance.
Source: Cybrarians Journal
Author: Ahmed Mokhtar
Institution: Minia University
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4405780346