Mosaicplots have been around now for quite some time, however, they are not very well understood. This may both be due to their high dimensionality as well as their hierarchical construction, which makes it hard to figure out what exactly can be seen in a picture as well as to decide which plot to draw given a specific problem. This paper will discuss these problems by giving yet another approach to mosaicplots, based on the underlying geometrical model. We will have a closer look at the popular statement that mosaicplots are the "graphical equivalent of contingency tables".This paper gives a wrap-up of mosaicplots in default construction mode as well as for some of their variations, such as fluctuation mode, Doubledecker plots and same-bin-size display.
[1]
Heike Hofmann,et al.
Generalized Odds Ratios for Visual Modeling
,
2001
.
[2]
Michael Friendly,et al.
Conceptual and Visual Models for Categorical Data
,
1995
.
[3]
W. Cleveland,et al.
Graphical Perception: Theory, Experimentation, and Application to the Development of Graphical Methods
,
1984
.
[4]
Ben Shneiderman,et al.
Tree visualization with tree-maps: 2-d space-filling approach
,
1992,
TOGS.
[5]
Martin Theus,et al.
Visualizing Loglinear Models
,
1999
.
[6]
J. A. Hartigan,et al.
Mosaics for Contingency Tables
,
1981
.