After-class self-evaluated comments from students are valuable for gaining information on their learning and reflection on the teacher’s instruction. Therefore, we developed a teacher dashboard system that automatically analyzes and visualizes students’ after-class self-evaluated comments based on the content aspect and sentiment of comments. The dashboard system allows students to input their after-class self-evaluated comments using two automatic classification mechanisms. The system is used to analyze the content aspect and sentiment categories of comments for each class and visualize the results. This helps teachers understand students’ learning statuses and decide whether adjustments or improvements to the course are necessary. These classification mechanisms include aspect-based and sentiment-based classifications. The aspect-based classification categorizes comments into six categories: the interestingness of teaching materials, the difficulty level of content, the learning gain of the class, the teacher’s teaching, other factors, and nonsense. The sentiment-based classification sorts student feedback into three sentiments: positive, neutral, and negative. The teacher dashboard system draws pie charts and tables with various background colors to display the distributions of different aspect and sentiment categories of comments. Additionally, we compared the accuracies of different combinations of classification models, considering both segmented and non-segmented data, as well as balanced and unbalanced data. The BERT model showed the highest accuracy (85.2%) for aspect classification with segmented and unbalanced data, while the LSTM + Attention model achieved the highest accuracy (87.7%) for sentiment classification with segmented data that was balanced through random oversampling.