agram:
Scattergram visualization of associations between all pairs of variables.
Description
Uses the association array generated by tap to construct
an array of association diagrams (pie charts and scatterplots).
Usage
agram(dataSet, method, ranking, order, ...)
Arguments
dataSet
the n x m data frame representing n samples of the m variables.
method
the method to use for linear association (included in the visualization)
as performed by cor,
one of "pearson", "spearman", "kendall". The default
is "pearson".
ranking
a logical specifying whether or not to rank the data points before
producing the scatter plots. The default is FALSE.
order
determines the variable ordering.
one of FALSE (the default for no re-grouping) or
"PC" (for principal components re-grouping)
or "OLO" (for optimal leaf ordering)
...
passed on to graphics routines ...
Value
Returns nothing.
Details
Displays associations between all pairs of variables
as a graphics array using pie-charts above the diagonal
depicting the proportion of the total association
that can be explained by a linear trend in blue,
and the remaining non-linear association in red.
Below the diagonal, scatter plots are shown and colored according to
proportional non-linearity (see nl3 returned by ma.nl ).
References
This code was adapted from the corrgram package:
Author: Kevin Wright,
Copyright: 2006 Kevin Wright,
License: GPL2.