This study considers inspection conducted in software development PBL as learning feedback and investigates the impact of each inspection comment on students. The authors have already collected most inspection comments for not only requirements specification but also UML diagrams on GitHub. The authors develop a tool that collects comments given in Figma to GitHub. We examine the impact on students of each classification of inspection comments based on the post-lesson questionnaire submitted by the students. Finally, we present the benefits that classification of inspection comments can bring to PBL and discuss automatic comment classification by machine learning enabled by text-based comments and the concept of software development PBL support application enabled by automatic classification of inspection comments.