pylab.plot() can take y as a matrix. When it does, it treats columns as data series. This makes it work nicely with, for example, pandas dataframes.

pylab.stackplot() takes y as a transposed matrix, where rows are data sereies, to data frames must be transposed. It would be better for consistency if it followed the plot format.

Also, plot() can take a y with no x, and x is generated as range(y.shape[0]) - it'd be great if stackplot allowed this as well.

ref: #819