Visualization x Artificial Intelligence
Visualization and AI increasingly shape one another: AI can power adaptive, efficient, and personalized visualizations, while visualization makes AI systems more interpretable, transparent, and human-centered. Our work explores these intersections, how machine learning can enhance visual design, how visualization can support AI literacy, and how visual tools can surface the social and ethical dimensions of data and AI practices. Through human-centered visual design, we aim to advance more inclusive, responsible, and value-driven uses of AI.
Color Crafting: Automating the Construction of Designer Quality Color Ramps IEEE VIS 2019
This project applies AI to automate color design in data visualization, enabling non-experts to create perceptually effective and aesthetically pleasant color ramps with ease. By analyzing a dataset of 222 expert-crafted color ramps, we trained a machine learning model to capture common patterns and design strategies. The system generates high-quality sequential and diverging color scales from a single seed color, reflecting principles of clarity, visual balance, and designer intuition. Crowdsourced evaluations show that the AI-generated ramps perform comparably to expert-designed ones, expanding access to thoughtful color design and supporting more expressive, accessible data visualizations.