<![CDATA[Text-based emotion classification results are commonly presented as labels, tables, or probability scores, making relationships among emotions difficult for users to interpret intuitively. This problem indicates the need for a visualization approach that not only presents the dominant emotion but also shows the intensity and relationship of supporting emotions in a single visual representation. This study aims to develop an interactive emotion visualization system for controlled test texts by integrating the RoBERTa GoEmotions model and the Plutchik Emotion Wheel. The system processes input text through light preprocessing, Transformer model inference, rule based emotion mapping into the eight basic Plutchik emotions, max normalization, visual threshold filtering, and radial visualization using D3.js. The contribution of this study lies in integrating the more detailed GoEmotions output into the basic emotion structure of Plutchik, applying normalization and visual threshold filtering to improve readability, and developing an interactive visualization that supports the interpretation of text emotion analysis results. Functional testing was conducted to examine the consistency of the inference process, emotion mapping, score normalization, and visualization generation. The test example shows that a sentence expressing success produces joy as the dominant emotion, with an aggregate score of 1.220 and a normalized value of 1.000. This result is consistent with the context of the input text and is visualized as the most dominant petal in the Plutchik Emotion Wheel. The findings indicate that integrating modern NLP models with interactive visualization can improve the interpretability of text-based emotion analysis compared with numerical or tabular presentation.]]>
