|
Abstract:
|
This talk investigates what influence the size of a dataset has on standard statistical graphics. Area plots---like bar charts, histograms, and mosaic plots---are relatively invariant against the increasing size of a dataset. Point plots, all plots that plot a glyph for each observation, have to cope with the problem of overplotting and an increasing number of extreme outliers. The definitions of other plots (e.g., boxplots) are examined to see if they still seem to be appropriate when datasets get large. Problems, solutions, and modifications are presented, which are based on either more interactivity or on advanced plotting techniques.
|
