Glyph Plot, Andrews Plot and Parallel Coordinates Plot

For multidimentional data (multivariate), which means data that has many attributes, visualizing data using these techniques might be useful. The input is a matrix where rows are observations (samples) and columns are variables (attributes).

GlyphPlot displays the data in flower-like icons, sometimes it is slightly different and called “starplot”, “roseplot” or “spiderplot”. For example:

>>load cereal;
>>glyphplot(cereal);

MATLAB glyphplot can show in glyph in Chernoff's “face” style too.

>>glyphplot(cereal,’glyph’,’face’);

Andrews Curves transform each observation into a function f(t) and plot the curve. It is useful in grouping the data.

>> load fisheriris
>> andrewsplot(meas,'group', species);

Parallel Coordinates Plot is similar to Andrews Curves but without the transformation.

>> labs = {'Sepal Length','Sepal Width','Petal Length','Petal Width'};
>> parallelcoords(meas, 'group',species);

More on visualizing multivariate data.