Dataset of User Reviews from Low-Rated Applications on the Amazon Appstore

This dataset provides a comprehensive collection of 79,821 end-user reviews from 64 low-rated software applications on the Amazon Appstore. The dataset is specifically curated to focus on applications that exhibit significant user dissatisfaction, making it a valuable resource for studying the root causes of software failure and negative user reception. The applications span 14 distinct categories, offering a broad view of user feedback across different software domains.

The data was systematically collected using an automated web scraping tool. The selection criteria targeted applications with a user rating of 3 stars (out of 5) or lower and a minimum of 400 reviews to ensure the data reflects a broad consensus of user opinion. The dataset is provided in CSV format and contains three primary columns for each review: the user's rating ('Stars'), the 'Title_of_Review', and the full text of the 'Base_Review'.

This resource is intended for researchers, software developers, requirements engineers, and educators. It is particularly useful for studies in software quality assessment, user experience (UX) analysis, sentiment analysis, issue detection, and for training and validating machine learning and natural language processing (NLP) models. Its unique focus on negative feedback provides an unfiltered and concentrated source of data for understanding why software products fail to meet user expectations.