Empowering Non-Technical Humanities Scholars with AI-Driven Tools: An Accessible Framework for Poetry Analysis

This paper presents an accessible, AI-powered framework designed to empower non-technical humanities scholars in emotion analysis and data visualization for poetry interpretation. By integrating state-of-the-art natural language processing (NLP) models—including BERT for emotion classification and GPT-based data visualization tools—the framework enables scholars to explore the psychological dimensions of poetic language, identifying anger, disgust, sadness, joy, neutrality, surprise, and fear with high accuracy. The workflow provides an accessible solution, eliminating technical barriers and allowing scholars to input text, extract emotions, and visualize findings without programming expertise. The system leverages BERT-based deep learning models to conduct fine-grained sentiment classification, capturing the complex emotional structures embedded in poetic texts. GPT-driven visualization tools, including word clouds and heatmaps of emotion correlations, translate these insights into intuitive, interpretable graphical representations. Through a case study of Emily Dickinson’s poetry, this framework demonstrates how AI can uncover latent patterns of grief, existential reflection, and mortality awareness, complementing traditional literary analysis with data-driven objectivity. This approach expands the methodological landscape of digital humanities, offering new avenues for creative engagement, interpretation, and interdisciplinary research. By democratizing access to AI-driven emotion analysis and visualization, this work paves the way for a new paradigm in human-AI collaboration, enabling scholars without technical expertise to critically engage with literature through computational perspectives while preserving the nuances of human interpretation.