Quantifying the Double Diamond AI-driven analysis of design thinking patterns in university classrooms
Publication Type
Conference Proceeding Article
Publication Date
8-2025
Abstract
The Double Diamond model is widely recognised in design education as a framework for structuring the design process, yet empirical evidence in student work remains limited. This study presents a novel approach to quantify the Double Diamond pattern using AI-driven text analysis of weekly student reflections collected from four design classes across two public universities in Singapore. Using Natural Language Processing (NLP), specifically Word2Vec embeddings and t-SNE visualization, we converted individual student reflections into quantifiable measures of divergent and convergent thinking over a 14-week semester. Through this approach, students progression can be quantified and visualised into a dashboardlike system, mapping individual progression using divergence and convergence as a proxy. To account for different pacing across classes, we analysed five critical phase transition points identified in each course’s syllabus instead of the original term-long based approach. Our findings revealed that 20 out of 22 teams exhibited a distinctive double-peak pattern in their divergence metrics, corresponding to the two divergent phases (Discover and Develop) of the Double Diamond model. This preliminary evidence suggests that the Double Diamond’s characteristic pattern can be quantitatively observed in student design processes, offering a new method for tracking design thinking progression at both individual and team levels.
Keywords
Data Analysis; Data Visualization; Design Education; Design Thinking; Double Diamond; New Product/ Process Innovation
Discipline
Artificial Intelligence and Robotics | Educational Assessment, Evaluation, and Research | Higher Education
Publication
Proceedings of the ASME 2025 International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, Anaheim, CA. August 17-20
Volume
4
First Page
1
Last Page
9
ISBN
9780791889244
Identifier
10.1115/DETC2025-165791
Publisher
American Society of Mechanical Engineers (ASME)
City or Country
Anaheim
Citation
CHIU, Matt; MAKANY, Tamas; and SILVA, Arlindo.
Quantifying the Double Diamond AI-driven analysis of design thinking patterns in university classrooms. (2025). Proceedings of the ASME 2025 International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, Anaheim, CA. August 17-20. 4, 1-9.
Available at: https://ink.library.smu.edu.sg/lkcsb_research/7869
Additional URL
https://doi.org/10.1115/DETC2025-165791