How to plot one curve and change color according to value

The tricky thing is that I get an attribute with the values and would like to have this part in the plot in a different color:

x = linspace(0,10);

y = sin(3*x).*exp(0.5*x);

c = find(y > 25);

m(length(x)) = 0;

m(c) = 1;

figure;

plot(x, y, 'b')

Where m becomes 1 the color should be red, else blue.

Can someone find the easiest way to do this?

[Manitux]

Ah, I see there is a misunderstanding. The "m" is part of the transmitted data, not a special limit. For the demo-code it was just generated as an example.

Answer 1

[DGM]

Also:

% inputs
thresh = 25;
x = linspace(0,10,100);
y = sin(3*x).*exp(0.5*x);

% m is equivalent to (y>thresh), so it's entirely redundant
% adding 1 here casts the logical result to float
% and converts it to a 1-based index vector for direct
% addressing of the color table (CT)
c = (y > thresh) + 1; 

% create an unfilled patch
hp = patch([x NaN],[y NaN],[c NaN],'EdgeColor','interp');
hp.LineWidth = 2; % if you want fatter lines
yline(thresh);

% apply the color
CT = [0 0 1; 1 0 0];
colormap(CT)

Of course, the color transitions where ever there are samples in the data, not necessarily exactly on y=25.

[Manitux]

Yes, this will do it. Just adjust a bit. Like this:

x = linspace(0,10);

y = sin(3*x).*exp(0.5*x);

c = find(y > 25);

m(length(x)) = 0;

m(c) = 1;

patch([x NaN],[y NaN],[m NaN],'EdgeColor','interp');

CT = [0 0 1; 1 0 0];

colormap(CT)

Thanks a lot.

[DGM]

As I mentioned, there's no need to.

x = linspace(0,10);
y = sin(3*x).*exp(0.5*x);

% you could do this one thing
c0 = y > 25;

% or you could do all this extra stuff
% to turn a mask back into a copy of itself
c = find(c0);
m(length(x)) = 0;
m(c) = 1;

% numerically, they're identical
isequal(double(c0),m)

ans = logical
   1

[Manitux]

The "m" was just a sample data (the curve as well). Actually both are part of the transmitted data, "y" is an analogue value and "m" a bool. In real it may look like this:

m([30 31 32 33 34 35 36 37 60 61 62 63 64 65 90 91 92 93 94 95]) = 1;

[Image Analyst]

I don't understand why we need both m and c and c0. Isn't it just y > 25, like

m = y > 25

[Manitux]

Try this:

x = linspace(0,10);

y = randi([-50 50], 1, length(x));

m(length(x)) = 0;

m([30 31 32 33 34 35 36 37 60 61 62 63 64 65 90 91 92 93 94 95]) = 1;

patch([x NaN],[y NaN],[m NaN],'EdgeColor','interp');

CT = [0 0 1; 1 0 0];

colormap(CT)

"y" and "m" are signals I receive on an interface and need to paint "y" in 2 different colors, depending on the bool of "m". My first sample was just how I tried to get this solved (by myself).

But your idea to take "patch" is brilliant ;-)

Answer 2

[Image Analyst]

Maybe this:

x = linspace(0,10);
y = sin(3*x) .* exp(0.5 * x);
plot(x,y,'-b.'); 
hold on
yline(25, 'LineWidth', 2, 'Color', 'm');
mask = y > 25;
y1 = nan(1, numel(y));
y1(mask) = y(mask);
plot(x, y1,'-r.');
grid on;

[William Rose]

@Image Analyst, nice.

Answer 3

[John D'Errico]

As ifs often the case, I am far too late to the party. :) But there are often many ways to solve a problem, so I like to be able to offer an alternative. scatter is one in this case.

x = 1:100;
y = sin(x/10);
scatter(x,y,[],y > 0.5)
yline(0.5,'r')

For more complex cases, scatter can still work. And, of course, we can control the colormap used.

k = cos(x/10) > 0;
scatter(x,y,[],k)
colormap([0 1 0;0 0 1])

And finally, scatter will alllow me to segregate multiple sections on the curve, according to my choosing.

k = round(cos(x/10));
scatter(x,y,[],k)
colormap([1 0 0;0 1 0;0 0 1])
colorbar

Answer 4

[William Rose]

@Fritz Weld,

[Edit: clean up code a lttle bit.]

x = linspace(0,10);
y = sin(3*x).*exp(0.5*x);
y1=zeros(size(y)); y2=y1;  % initialize
for i=1:length(y)
    if y(i)<=25
        y1(i)=y(i);
        y2(i)=NaN;
    else
        y1(i)=NaN;
        y2(i)=y(i);
    end
end
plot(x,y1,'-b.',x,y2,'-r.');

Probably not the easiest or prettiest way to do it but it works.

Another approach is to use scatter with a colormap.

[William Rose]

@Fritz Weld,

x = linspace(0,10);
y = sin(3*x).*exp(0.5*x);
c=(y>25);
map=[0,0,1;1,0,0];
scatter(x,y,[],c,'filled');
colormap(gca,map);

"scatter" will not connect the points.