Hello!
Welcome to my website. My name is Daye Kang :)
My research interests lie at the intersection of Human-Computer Interaction (HCI), human-AI collaboration in data analysis and visual communication. I design and study technologies to promote data sensemaking, creativity, and well-being. Recently, I have been designing an intelligent system that utilizes LLMs and interactive data visualization to assist qualitative data analysis.
I intend to apply such interactions to assist AI non-experts, such as patients, in analyzing and understanding their health data to promote self-reflection. I am interested in developing novel visual media that can deliver qualitative and quantitative data together to promote comprehensive and engaging visual storytelling.
I started my academic career in Visual Communication Design, focusing on UX research and animation at Hongik University. I received my MS in Industrial Design, focusing on HCI, from KAIST, advised by Prof.Andrea Bianchi. Currently, I am a Ph.D. student at Cornell University's Information Science Department, working with my advisor, Prof.Jeff Rzeszotarski, and many wonderful colleagues.
You can pronounce my name as Da-Ye [da ye].
News
-
[May'24] Sharing 'LLM-embedded interactive visual system for iterative theme refinement' at CHI2024 workshop. LLMs as Research Tools: Applications and Evaluations in HCI Data Work
-
[March'24] The 'Challenges and Opportunities for Tool Adoption in Industrial UX Research Collaborations' paper was accepted to CSCW'24
RESEARCH
Publications
CSCW 24
Challenges and Opportunities for Tool Adoption in Industrial UX Research Collaborations
Daye Kang, and Jeff Rzeszotarski
Qualitative UX research practitioners analyze data to comprehend users' needs and synthesize implications for future design. Working with multiple stakeholders is inevitable in modern product development, and in industry, coordination and collaboration add pressures to already laborious qualitative data analysis tasks. In our work, we investigate how multi-stakeholder collaboration specifically affects industry qualitative data analysis practices. We investigate the challenges practitioners face, limitations of current tools, and opportunities for computer-supported cooperative work.
[PDF (Draft)] | [To be appeared]
CHI 2021
ToonNote: Improving Communication in
Computational Notebooks Using
Interactive Data Comics
Daye Kang, Tony Ho, Nicolai Marquardt, Bilge Mutlu, Andrea Bianchi
ToonNote is a novel technique for representing computational notebooks in the form of interactive data comics. ToonNote provides a high-level, curated narrative of the dataset in Comic View. Readers can focus on data storytelling, and not be hindered by code, unnecessary outputs, or markdown and can switch back to the Notebook View when needed.
PDF | GitHub
IMWUI 2019
HeyTeddy: Conversational Test-Driven Development for Physical Computing
Yoonji Kim, Youngkyung Choi, Daye Kang, Minkyeong Lee, Tek-Jin Nam, Andrea Bianchi
Physical computing is a complex activity that consists of different but tightly coupled tasks: programming and assembling hardware for circuits. We propose a general-purpose prototyping tool based on conversation. HeyTeddy guides users during hardware assembly by providing additional information on requests or by interactively presenting the assembly steps to build a circuit.
DIS 2019
Designing Internal Structure of Chocolate and Its Effect on Food Texture
Yujin Lee, Jee Bin Yim, Daye Kang, Hyeon-Beom Yi, Daniel Saakes
In this research, we explored the effects of the internal structure on the texture of chocolate by designing, building, and testing chocolates with different internal structures and internal chocolate percentage. Multiple layers of patterned chocolate were stacked as a fabrication method. Each layer was fabricated using a silicone mold that was made using a 3D printed model. We created seven types of chocolate variations through this system.
KSDS 2017 (Korean Society of Design Science)
Lexia in Wonderland: Korean Education Service for Children with Dyslexia
Daye Kang, Hye-Ryeong Kim, Ji-Hae Lee, Jae Young Yun.
Treating dyslexia during childhood is very important. In Korea, this is not the case since the condition is not well understood. Also, there is a limited number of medical centers that offer treatment for it. We propose an education app for dyslexic children ages 5-10 in Korea that can help these children through active learning. By teaching the children Korean phonics using 'nonwords' in a fun and engaging way, they can be treated at home instead of visiting the medical centers.
KSDS 2017 (Korean Society of Design Science)
Nudge Design to Increase Physical Activities for Hospitalized Children
Daye Kang, Hye-Min Choi, Ka-Hyun Kim, Younjoon Lee.
Hospitalized children are often asked to regularly walk in the hallway to stay active. However, the hallway is usually dull and empty. We created wallpapers and floor footprint stickers that use nudge design to be fun and engaging. Footprint stickers promote walking and leg stretches while the wallpaper contains stories that can add fun moments. Children are guided through stretching activities indicated on the wall.