change colors of a scatter plot with double y axis

Question

I've searched around, but have not found a solution. My goal is to plot a scatter plot with 3 sets of data, each set with a different color. Here is an example of my code:

%generate input
x1=[732490 732509 732512 732513 732521 732528];
y1=[7.828 7.609 22.422 14.758 26.258 1.477];
x2=[732402 732403 732404 732404 732433 732555];
y2=[0.693 0.645 0.668 0.669 0.668 0.662];
x3=[832402 832403 832404 832404 832433 835423];
y3=[0.693 0.645 0.668 0.669 0.668 0.685];
figure(1); 
[ax,h1,h2]=plotyy(x1,y1,[x2,x3],[y2,y3],'scatter');
blue=[0 0 1];
red=[1 0 0];
set(h1,'cdata',red);
set(h2,'cdata',blue);
set(ax(1),'ycolor','r');
set(ax(2),'ycolor','b');

However, this is what exactly what I want, as [x2 y2] [x3 y3] has the same color. Is there a way to change the colors so that the three set of data has different colors? And also how to add a legend showing the three sets of data?

Answer 1

You can overcome the limitation of plotyy by building your own plotyy by:

• create a figure
• add an axes
• add a second axes and:
  • make its position equal to th eone of the first axes
  • make it transparent by setting its color to `none"

Now you can select the axes on which to use scatter by specifying it as a first argument (e. g. scatter(axes_1, ...)).

After having plot all the data set, you have to make the xlim of the two axes equals.

To add the legend, you just have to specify the handles" returned by thescatterfunction as first argument of thelegend` function.

This approach has been implemented in the following code.

In the code a check is made in order to verify if it is possible to use the dot notation (introduced in R2014b).

x1=[732490 732509 732512 732513 732521 732528];
y1=[7.828 7.609 22.422 14.758 26.258 1.477];
x2=[732402 732403 732404 732404 732433 732555];
y2=[0.693 0.645 0.668 0.669 0.668 0.662];
x3=[832402 832403 832404 832404 832433 835423];
y3=[0.693 0.645 0.668 0.669 0.668 0.685];
% Check if the "dot notation" can be used
dot_notation=~verLessThan('matlab','8.4')
%
figure
% Add the first axes
ax1=axes
% Add the second axes
ax2=axes
% Plot the scatter of the first set of data on the first axes
h1=scatter(ax1,x1,y1,'r','filled')
% Plot the scatter of the second set of data on the second axes
h2=scatter(ax2,x2,y2,'b','filled')
hold on
% Plot the scatter of the third set of data on the second axes
h3=scatter(ax2,x3,y3,'g','filled')
if(dot_notation)
   % Superimpose the second axes over the first ome
   ax2.Position=ax1.Position
   % Make it transparent
   ax2.Color='none'
   % Move the YAxis to the right
   ax2.YAxisLocation='right'
   % Adjust the X limits
   x_lim=[min([ax1.XLim ax2.XLim]) max([ax1.XLim ax2.XLim])]
   ax1.XLim=x_lim
   ax2.XLim=x_lim
   % Remove XAxis Tick
   ax2.XTick=[]
else
   ax1_pos=get(ax1,'position');
   % Superimpose the second axes over the first ome
   set(ax2,'Position',ax1_pos)
   % Make it transparent
   set(ax2,'color','none')
   % Move the YAxis to the right
   set(ax2,'YAxisLocation','right')
   % Adjust the X limits
   ax1_x_lim=get(ax1,'xLim');
   ax2_x_lim=get(ax2,'xLim');
   x_lim=[min([ax1_x_lim ax2_x_lim]) max([ax1_x_lim ax2_x_lim])]
   set(ax1,'XLim',x_lim)
   set(ax2,'XLim',x_lim)
end
% Add the legend
[a,b,c,d]=legend(ax2,[h1,h2,h3],'1° data set','2° Data set','3° Data set')
if(dot_notation)
   a.Color='none'
else
   set(a,'color','w')
end
grid(ax1,'on')

Hope this helps.

Qapla'

Answer 2

In this example from matlab it is done without the scatter:

x = linspace(0,10);
y1 = 200*exp(-0.05*x).*sin(x);
y2 = 0.8*exp(-0.5*x).*sin(10*x);
y3 = 0.2*exp(-0.5*x).*sin(10*x);

figure
[hAx,hLine1,hLine2] = plotyy(x,y1,[x',x'],[y2',y3']);

notice that if instead you write [hAx,hLine1,hLine2] = plotyy(x,y1,[x,x],[y2,y3]); you'll loose the third color, so you need the [x',x'] instead of the [x,x]

I just tested your code but with [ax,h1,h2]=plotyy(x1,y1,[x2',x3'],[y2',y3']); and I could see the 3 colors. Sadly when I do [ax,h1,h2]=plotyy(x1,y1,[x2',x3'],[y2',y3'],'scatter'); I keep getting the same

Error using scatter (line 44)
X and Y must be vectors of the same length

, even when I tried on the code from the matlab example. Apparently the scatter property does not allow you to have 3 set of data. If you check line 42 of scatter.m you'll see

[~, cax, args] = parseplotapi(varargin{:},'-mfilename',mfilename);

When you execute [hAx,hLine1,hLine2] = plotyy(x,y1,x,y2,'scatter'); which is only a 2 set plot you'll see how args in line 42 is a 1x2 cell with two vectors, yet when you execute [ax,h1,h2]=plotyy(x1,y1,[x2',x3'],[y2',y3'],'scatter'); which would be the only way of getting 3 colors

(remember [ax,h1,h2]=plotyy(x1,y1,[x2,x3],[y2,y3],'scatter') will only give you 2 colors although it works within scatter.m as in args that is also interpreted as 2 vectors)

as now args in line 42 is a 1x2 cell with two matrices instead of two vectors that yields to a mistake.

Moreover using:

blue=[0 0 1];
red=[1 0 0];
set(h1,'cdata',red);
set(h2,'cdata',blue);
set(ax(1),'ycolor','r');
set(ax(2),'ycolor','b');

Isn't helpful as you can only manipulate 2 instead of three independent axis. So I guess that the answer to your question is NO IT CAN'T BE DONE (Although if you remove the scatter constraint it will).