Abstract

With the growing volume of textual data across the various fields within design disciplines, there is an opportunity to analyze and visualize trends within such data using advanced methods. Traditional techniques, such as word clouds, offer a simplistic overview of textual content and often lack the statistical rigor necessary for deeper insights. This study explores topic modeling as a viable advanced method for analyzing latent patterns in textual data. Through a case study of Design Journal articles published between 2021 and 2023, we demonstrate how topic modeling, specifically BERTopic, can uncover research trends and provide more detailed and insightful visualizations compared to those that rely solely on word-frequency-based methods. By integrating techniques like HDBSCAN (Hierarchical Density-Based Spatial Clustering of Applications with Noise) for clustering within BERTopic, we emphasize the importance of fine-tuning the model’s parameters to extract relevant topics from design research articles from previous years. The findings reveal that design research is diverse in its breadth and depth of topics; from conducting research in typeface legibility to behavior economics, researchers in the broad field of design are constantly expanding the edges of their domains by pursuing new approaches and addressing a wide range of relevant real-world challenges.