define color range in scatter plot

Hi,

I have values ranging from 0 to 3000. Here, only for the values ranging between 2000 to 3000 I have to use the color palette, for rest of the values (i.e. 0 to 1000) the color should be in black..

And I am using scatter plot here.. Is it possible to define above conditions in the scatter plot..

Answer 1

[Jonas]

i do not fully understand, you want to use the pallett for split to 0 to 3000, but at a certain threshold you want to use black instead?

what about values between 1000 and 2000?

x = linspace(0,3*pi,2000);
y = cos(x) + rand(1,2000);
c = linspace(1,10,length(x));
sp=scatter(x,y,[],c);

currMap=colormap();

[~,highestCDataIdx]=max(sp.CData(sp.XData<=3));

highestCDataIdx=ceil(highestCDataIdx*size(currMap,1)/numel(sp.XData)); % scale CData Index by ration of colormap elements to number of XData to map

currMap(1:highestCDataIdx,:)=0;

colormap(currMap)

you can also change values inbetween without changing the pallett

x = linspace(0,3*pi,2000);
y = cos(x) + rand(1,2000);
c = linspace(1,10,length(x));
sp=scatter(x,y,[],c);

currMap=colormap();

[~,highestCDataIdx]=max(sp.CData(sp.XData<=3));

highestCDataIdx=ceil(highestCDataIdx*size(currMap,1)/numel(sp.XData)); % scale CData Index by ration of colormap elements to number of XData to map

currMap(1:highestCDataIdx,:)=0;

colormap(currMap);


delInbetween=sp.XData<5 & sp.XData>4;

sp.XData(delInbetween)=[];
sp.YData(delInbetween)=[];
sp.CData(delInbetween)=[];

[Turbulence Analysis]

Thanks, Jonas..

This is what I am exactly looking at..

Actually,the values between 0 to 2000 should be in black and palette for the rest i.e. 2000 to 3000

[Jonas]

so the full pallett for the rest or just parts of the palette?

or are you already happy with the given code?

clear

x = linspace(0,3000,20000);
y = 3*sin(2*pi*1/1000*x)+rand(1,20000);
c = linspace(1,10,length(x));

figure;
scatter(x,y,[],c);

figure;
sp=scatter(x,y,[],c);

currMap=colormap();

[~,highestCDataIdx]=max(sp.CData(sp.XData<=2000));

highestCDataIdx=ceil(highestCDataIdx*size(currMap,1)/numel(sp.XData)); % scale CData Index by ration of colormap elements to number of XData to map

currMap(1:highestCDataIdx,:)=0;
colormap(currMap);

figure;
y=y(x>2000);
x=x(x>2000);
c = linspace(1,10,length(x));

scatter(x,y,[],c);
xlim([0 3000])

linkaxes()

[Turbulence Analysis]

Thank you very much!!

[Turbulence Analysis]

Jonas,

Is there a way to fix particular color rather than a gradient , lets say

for 0 to 2 - green;

2 to 4 - red;

4 to 6 - yellow

something like this

Thanks !!

[Jonas]

you can give each point its inidividual color, e.g.

clear
x = linspace(0,3,1000);
y = 3*sin(2*pi*1*x)+rand(1,1000);

c=zeros(numel(x),3);

firstColIdx=x<=1;
secColIdx=x<=2 & x>1;
thirdColIdx=x<=3 & x>2;

firstColor=[1 0.5 0];
secondColor=[0.5 0 0.5];
thirdColor=[0 1 0];


c(firstColIdx,:)=repmat(firstColor,[sum(firstColIdx),1]);
c(secColIdx,:)=repmat(secondColor,[sum(secColIdx),1]);
c(thirdColIdx,:)=repmat(thirdColor,[sum(thirdColIdx),1]);


figure;
scatter(x,y,[],c);

[Turbulence Analysis]

Jonas, actually in my case the varible c is not defined based on x, its rather different as shown in the herewith attachment..

[Jonas]

i got a matrix with 4 columns. how to process this, how can i see what you want?

if you plot the 4 lines only with plot(yourMatrix), then your x is 1:size(yourMatrix,1)

[Turbulence Analysis]

Right now actually I am plotting like this.

As you can see I coloring based on entries in the matrix c. Since the selected colormap adopts gradient scale sometimes it is very difficult to differentiate..

figure ()
ax = gca;
scatter(x(:,1),y(:,1),0.2,c(:,1))
ylim ([0 3500]);
xlim ([0 1]);
%grid on
set(gca,'TickLabelInterpreter','latex')
set(gca, 'FontSize', 12)
title(sprintf("${Y=%3dmm}$",h),'FontSize',15,'Color','black','Interpreter','latex');
width=400;
height=300;
ax.LineWidth = 2;
set(gcf,'position',[x0,y0,width,height])
colormap (flipud(colorcube))
c2 = colorbar('FontSize',12,'TickLabelInterpreter', 'latex');
caxis ([-10 35]);

[Jonas]

and how is the provided matrix connected to your code? i neither have your x, y and c nor does the provided 4 column matrix occur in your code

so you use the colormap colorcube which i do not have and how do you want to see the given plot?

[Turbulence Analysis]

Ah, Sorry.

I dont know why my previous attachment does not contains the x y c. You can find the same in this new attachment..

Actually I dont have any restriction with using colorcube, All I am looking for is colormap without gradient apparently colorcube is not.. So I am looking for another options..

The one you showed this morning is really promising, but I am having some while using my data !!

[Jonas]

so some things to make the points in your cloud better distinctable: you can change the marker and the marker size and the coarseness of you colormap:

load('matlab.mat','x','y')
scatter(x,y,[],linspace(0,1,numel(x)),'.','SizeData',1)
colorbar;

cm=colormap('colorcube');
cm=cm(1:8:end,:);
colormap(cm)

or, if you want to use the supplied c for whatever reason

load('matlab.mat','c');
figure;
scatter(x,y,[],c,'.','SizeData',1)
colorbar;

cm=colormap('colorcube');
cm=cm(1:8:end,:);
colormap(cm)

[Turbulence Analysis]

Thank you very much, Jonas!!

Answer 2

[KSSV]

Read about clim, caxis.