Reputation: 1511
How can I draw a line whose thickness varies at each point in Matlab? I need to plot an average line, and then the standard deviation plotted as a shadow below it. Any ideas?
Thanks,
Upvotes: 1
Views: 1010
Reputation: 814
This is something I've written a while ago - it's a bit long, but you can just copy the whole thing. It deals with variable size input and produces a beautifully shaded plot - with std and min/max. Signal should be 2D. If you need a version for plotting two signals - just ask :)
function H = plotp(varargin)
switch nargin
case{1}
signal = varargin{1};
time = 1:size(signal,2);
prop = 'r-';
new_figure = true;
case{2}
time = varargin{1};
signal = varargin{2};
prop = 'r-';
new_figure = true;
case{3}
time = varargin{1};
signal = varargin{2};
prop = varargin{3};
new_figure = true;
case{4}
time = varargin{1};
signal = varargin{2};
prop = varargin{3};
new_figure = varargin{4};
end
if ischar(new_figure)
temp7 = regexpi(new_figure,'true');
if isempty(temp7)
new_figure = false;
H = get(0,'CurrentFigure');
else
new_figure = true;
end
elseif isnumeric(new_figure)
H = new_figure;
new_figure = false;
end
% prepare vectors for plotting
time2 = [time fliplr(time)];
sigm = nanmean(signal);
sigs = nanstd(signal);
sigms = [sigm-sigs fliplr(sigm+sigs)];
sigmin = nanmin(signal);
sigmax = nanmax(signal);
sigmm = [sigmin fliplr(sigmax)];
% check color
if strcmpi(prop(1),'r')
c1 = [1 0 0];
elseif strcmpi(prop(1),'g')
c1 = [0 1 0];
elseif strcmpi(prop(1),'b')
c1 = [0 0 1];
else
c1 = [1 1 1];
end
color1 = c1 + .7*(1-c1);
color2 = c1 + .8*(1-c1);
%
if length(prop) == 1
prop(2) = '-';
end
if new_figure
H = figure;
else
figure(H);
end
whole_screen = get(0,'ScreenSize');
% max figure size - add
fig_size = whole_screen + [-4 -4+2*32 +4+4 4+4-2*32];
set(H,'OuterPosition',fig_size);
plot(time,sigm,prop,'LineWidth',1.5)
hold all
fill(time2(~isnan(sigms)), sigms(~isnan(sigms)),color1,'EdgeColor',color1,'FaceAlpha', 0.4);
fill(time2(~isnan(sigmm)), sigmm(~isnan(sigmm)),color2,'EdgeColor',color2,'FaceAlpha', 0.3);
legend([{'Mean'} {'\pm Stddev'} {'Min/Max'} ],'Location','Best')
Upvotes: 2
Reputation: 53
could you plot three lines, one as an average and the other two as + and - standard deviations.
If you're feeling particularly masochistic you could calculate the confidence interval of your line and plot that.
Upvotes: 0