사용자 경험 문제 발견을 위한 생성형 AI 기반 앱 리뷰 데이터 분석 도구 개발 및 유용성 검증 = Development and Verification of A Generative AI Based App Review Analysis Tool for Discovering User Experience Issues

다국어 초록 (Multilingual Abstract)

Background: App reviews are qualitative data in which users freely describe functional issues, usability complaints, requirements, and emotional responses based on their actual experiences in the mobile environment. These reviews are valuable resources for improving user experience, but their unstructured and large-scale nature makes them difficult to analyze, especially for novice designers who lack experience with app review data analysis. This study aimed to develop a tool that enables such designers to leverage generative AI to more easily analyze app review data and explore user experience problems.
Methods: The study was conducted in four stages, focusing on the development and evaluation of a generative AI–based app review analysis tool. First, relevant prior studies on app reviews and user experience exploration were reviewed to derive analysis functions. Second, based on these functions, the tool was designed using ChatGPT. Third, two workshops were conducted with design students to test the applicability of the tool. Finally, the workshop activities and post-evaluation results were analyzed to assess the tool’s usefulness.
Results: The tool proved to be an effective means for novice designers to quickly and easily identify user experience problems from app reviews. Through the workshops, participants also learned new approaches to data analysis. In particular, they practiced developing perspectives for interpreting app review data and formulating questions that lead to meaningful insights.
Conclusion: This study developed a generative AI–based tool for exploring user experience problems from app review data and validated its applicability through workshops. The findings indicate that the tool improves both the speed and structure of review analysis, while also having positive effects on learning and research design processes. However, limitations remain regarding reliability, data scope, and the lack of quantitative evaluation. Future research should therefore examine broader service domains and practical applications of the tool.