Shaded plots in matlab

When comparing different traces in a plot, the usual approach is to plot the traces in different colors and add the confidence bounds as thinner lines. This works ok for two colors, but it ain’t the prettiest.

Just lines – acceptable, but not great

Luckily, its pretty easy to plot confidence bounds as filled patches. By making the patches transparent (alpha(x) in matlab), plots become much more manageable.

Patches&alpha make for prettier figures

This way, even large collections of traces can be made usable

Many conditions, one plot

With a very simple wrapper function, this trick is as easy to use as plot():

function varargout = plotshaded(x,y,fstr);
% x: x coordinates
% y: either just one y vector, or 2xN or 3xN matrix of y-data
% fstr: format ('r' or 'b--' etc)
% example
% x=[-10:.1:10];plotshaded(x,[sin(x.*1.1)+1;sin(x*.9)-1],'r');

if size(y,1)>size(y,2)

if size(y,1)==1 % just plot one line

if size(y,1)==2 %plot shaded area
    px=[x,fliplr(x)]; % make closed patch
    py=[y(1,:), fliplr(y(2,:))];

if size(y,1)==3 % also draw mean
    py=[y(1,:), fliplr(y(3,:))];

alpha(.2); % make patch transparent

The only issue with this method is that at the time matlab can’t export figures with transparency to illustrator. A solution is to use Juerg Schwizer’s svg export function.

This entry was posted in Matlab. Bookmark the permalink.

Comments are closed.