Analysing user reviews of interactive educational apps: a sentiment analysis approach

• Published in: Interactive learning environments Vol. 32; no. 1; pp. 355 - 372
• Main Authors: Mondal, Aadi Swadipto, Zhu, Yuang, Bhagat, Kaushal Kumar, Giacaman, Nasser
• Format: Journal Article
• Language: English
• Published: Abingdon Routledge 02.01.2024; Taylor & Francis Ltd
• Subjects: Applications programs; Augmented reality; Computer Simulation; Data mining; deep learning; Designers; Education; educational apps; Educational Objectives; Labels; New technology; Public Opinion; Sentiment analysis; Software development; Usability; user reviews; User satisfaction; Virtual reality
• ISSN: 1049-4820; 1744-5191
• DOI: 10.1080/10494820.2022.2086578

The aim of this study is to investigate the public opinion (i.e., the learners themselves) of apps incorporating emerging technologies Augmented Reality (AR) and Virtual Reality (VR) and find out how satisfied mobile platform users are with these technologies for educational purposes. We performed sentiment analysis of more than one million reviews from 800 different Android apps collected by systematic scrapping of the Google Play Store. Identified apps were separated into five categories: AR, AR+Educational, VR, VR+Educational, and Educational. We identify reviews from each app as positive, negative, or neutral, and apply aspect labels depending on the content of the review. We adopted Bidirectional Encoder Representations from Transformers (BERT) and classify the labels into seven categories: technical issues, usability, content, user interaction, feature request, learning qualities, and advert-related. The results indicate that the positive sentiments are 54.6% for AR, 49.6% for AR+Educational, 47.7% for VR, 71.4% for VR+Educational, and 75.2% for Educational. The results suggest that education apps that do not incorporate AR or VR are receiving higher user satisfaction than apps that incorporate these emerging technologies. Analyzing and understanding user reviews will help instructional designers, software developers, and hardware designers to resolve the key inhibitors .of these apps.