In Progress
Beltran, J. A., Mofid, H., Parikh, H., Gondaliya, J., Guzman, D., Shah, J., Escobedo, L., & Cibrian, F. (2026).
Exploring manifold-based clustering techniques for enhanced inductive thematic analysis.
SEET—Software Engineering for Emerging Technologies, CCIS 2725, 464–480. Springer.
Escobedo, L., Cibrian, F. L., & Beltran, J. A. (2026).
The Learning Experience of Using LLM for Qualitative Research Methods Courses in HCI.
Human-Computer Interaction, HCII 2026, LNCS 16703, 75–86. Springer.
Cibrian, F. L., Beltran, J. A., & Escobedo, L. (2026).
The Hidden Curriculum of LLM-Assisted Qualitative Analysis in HCI Education.
Proceedings of the 8th Annual Symposium on HCI Education, 1–8. ACM.
In progress
ATA-LLM combines large language models, manifold learning, and clustering to support inductive thematic analysis. The framework helps researchers generate initial codes, explore relationships among codes, and develop themes while retaining human review and interpretation.
Open coding: an LLM proposes initial codes from qualitative text.
Code representation: embedding models convert codes into numerical vectors.
Visualization and clustering: manifold learning and clustering help researchers explore patterns among codes.
Theme development: researchers review code groupings and LLM-assisted theme suggestions against the source data.
The connected studies examine the technical framework, students’ experiences using ATA-LLM in a qualitative methods course, and the implications of LLM-assisted analysis for rigor, privacy, and responsibility in HCI education.
Franceli Cibrian
Assistant Professor
Electrical Engineering and Computer Science, Fowler School of Engineering
Lizbeth Escobedo
Assistant Professor
Faculty of Computer Science, Dalhousie University