Scatter chart with colours from dark to light along the x-axis and blue to red along the y-axis

Hi,

I made a scatter based on interpolation of RGB colours (using only R and B) based on the y-axis data. Now I want the colours to change from dark to light based on the data of x axis. How can I do that? Thanks in advance!

Here is what I have right now and what I get:

b = [0 0 1]; % Blue

r = [1 0 0]; % Red

x =[27.5 26.8 68 70.5 43.5 45.2 72];

y = [7.2 3 7.9 1.7 4.23 1.057 0.89];

y_min_max = [min(y);max(y)];

color = interp1(y_min_max,[b;r],y);

figure()

s = scatter(x,y,80,color,'filled');

grid on; grid minor; box on

xlabel('x')

ylabel('y')

Answer 1

[William Rose]

@T,

b = [0 0 1]; % Blue
r = [1 0 0]; % Red
%x =[27.5 26.8 68 70.5 43.5 45.2 72]; 
%y = [7.2 3 7.9 1.7 4.23 1.057 0.89]; 
x = rand(1,1000);
y = rand(1,1000);
y_min_max = [min(y);max(y)];
color = interp1(y_min_max,[b;r],y);         %initial color based on y value
lightness=(x-min(x))/(max(x)-min(x));  %lightness of each point, in [0,1]
lightness=repmat(lightness',1,3);
color2=color.*lightness;                %adjust lightness of each point
figure()
s = scatter(x,y,80,color2,'filled');
grid on; grid minor; box on 
xlabel('x')
ylabel('y')

Good luck.

[William Rose]

@T,

I think your use of interp1() in your original post is very clever.

Here is a version that uses the full range of colors and does not use interp1.

In this version I use the fact that the "value" argument in the HSV version of a color is the lightness. This simplifies the implementation of lightness.

x = rand(1,1000);
y = rand(1,1000);
N=length(x);                        %number of points
hue=(y-min(y))/(max(y)-min(y));     %hue   based on y value, in [0,1]
value=(x-min(x))/(max(x)-min(x));   %value based on x value, in [0,1]
c_hsv=[hue',ones(N,1),value'];      %colors specified as HSV, where S=1 for every point
color = hsv2rgb(c_hsv);             %convert color from hsv to rgb
figure()
s = scatter(x,y,80,color,'filled');
grid on

[William Rose]

@T,

If you want to restrict the color range to B to R, while using the approach above, do the following. This relies on the fact that hues from B to R correspond to the range 0.667 to 1.00 on the hue scale. Therefore the lowest y value gets mapped to hue=0.667 (i.e. blue) and the highest y value gets mapped to hue=1.00 (i.e. red).

x = rand(1,1000);
y = rand(1,1000);
N=length(x);                        %number of points
hue=2/3+(1/3)*(y-min(y))/(max(y)-min(y));  %hue   based on y value, in [0,1/3]
value=(x-min(x))/(max(x)-min(x));   %value based on x value, in [0,1]
c_hsv=[hue',ones(N,1),value'];      %colors specified as HSV, where S=1 for every point
color = hsv2rgb(c_hsv);             %convert color from hsv to rgb
figure()
s = scatter(x,y,80,color,'filled');
grid on

[T]

Thanks William!

[William Rose]

@T, you're welcome. The answer I posted as a comment is better, in my opinion, than my initial answer. Good luck.