Towards Understanding Human Similarity Perception in the Analysis of Large Sets of Scatter Plots