VisiCube is no longer supported:
for a more advanced version of the same product!
VisiCube's Scatter Matrix
A two-dimensional scatter plot is a bivariate plot in which
a comparison of 2 measures is presented, one measure along each axis.
A scatter plot provides a useful exploratory method for understanding
the correlation between the two measures
(which may lead to conclusions about the dependence of one measure on the other).
The following is a matrix of such plots. In a VisiCube matrix, you are
presented individual plots for a single set of records...each for a different measure.
A matrix gives you the ability to quickly compare the different measures.
In this matrix of iris data, where each of three species is
shown with a different color, you can compare measures as well
as the species being measured. With regards to measure comparisons, for example,
it can be seen that petal length and width correlete well (as one increases, so
does the other). On the other hand, sepal length and width do not correlate well
at all. It can also be seen that the petals of the Setosa species (here shown
in red) are consistently and markedly smaller than those of the Versicolor species
(shown in blue) and that Versicolor petals are smaller than Virginica petals
(shown in green). Finally, regarding sepal size, it appears that the Setosa sepals
are both shorter and wider than the other species (which themselves do not
differ much at all).
The data presented here is from E. Anderson's study of 3 species of iris,
The Irises of the Gaspe Peninsula, which was published in 1935 in the
Bulletin of the American Iris Society.
This data (and visual) is supplied with VisiCube as a sample project.