Abstract:
|
Bill Cleveland has made huge contributions to statistical graphics. In a paper he published with Robert McGill in 1984, Cleveland describes the results of an experiment that he conducted to determine the effectiveness of mapping data elements to visual elements in statistical graphics. They prepared a large packet of statistical plots with associated questions, and took these around the corridors of AT&T Bell Labs and home to their spouses to complete. The results indicated a hierarchy of the graphical elements based on the accuracy with which the reader returned the data values from the plot, which from best to worst are position along a common scale, position along nonaligned scales, length/direction/angle, area, volume/curvature, shading/color saturation. The experiment was comprehensively repeated in Heer and Bostock (2010) using subjects from Amazon's Mechanical Turk, and produced corroborating results. In this talk, to honor Bill's work, we will use interactive crowd-sourcing with the audience to test some of his cognitive principles.
|