A tableplot (Kwan, 2008) is designed as a semi-graphic display in the form of a table with numeric values, but supplemented by symbols with size proportional to cell value(s), and with other visual attributes (shape, color fill, background fill, etc.) that can be used to encode other information essential to direct visual understanding. Three-way arrays, where the last dimension corresponds to levels of a factor for which the first two dimensions are to be compared are handled by superimposing symbols.
The general graphic method was first designed as a visualization method of for exploratory and confirmatory factor analysis results (Kwan, 2008), allowing easy graphic comparison of alternative solutions, rotations, etc. Friendly & Kwan (2009) use this form for a new display of collinearity diagnostics, and Friendly & Kwan (2011) discuss the use of tableplots in a debate on the roles of tables and graphs in statistical presentation.
The main function is
provided for collinearity diagnostics (Friendly & Kwan, 2009).
In future releases, tableplot methods will be provided
for other specialized tables.
Ernest Kwan and Michael Friendly
Maintainer: Michael Friendly <email@example.com>
Kwan, E. (2008). Improving Factor Analysis in Psychology: Innovations Based on the Null Hypothesis Significance Testing Controversy. Ph. D. thesis, York University.
Friendly, M. & Kwan, E. (2009). Where's Waldo: Visualizing Collinearity Diagnostics, The American Statistician, 63(1), 56-65.
Kwan, E. and Lu, I. R. R. and Friendly, M. (2009). Tableplot: A new tool for assessing precise predictions, Zeitschrift für Psychologie / Journal of Psychology, 217, 38-48.
Friendly, M. & Kwan, E. (2011). Comment (Graph people versus table people: Reply to Gelman), Journal of Computational and Graphical Statistics, 20 (1), 18-27.
corrgram for corrgrams of correlation matrices
balloonplot for balloonplots of two-way tables
# See: demo(psych9)