Color Palette Generation for Nominal Encodings

Encoding data using color is a key technique employed by almost all visualizations. A set of colors used to represent data values must be carefully chosen in order to maximize the effectiveness of the visualization in which it is used. We present a method of automat- ically generating effective color palettes for nominal data encoding based on user specifications. We apply simulated annealing with an appropriate heuristic function to find palettes that have visually distinguishable colors and adhere closely to a user's color prefer- ences. Such palettes can be customized to feature color harmony, a particular variation of contrast or saturation, or support for the colorblind.