How to plot the correct axes for a heatmap

I have a set of 2D coordinate data for 5598 points (attached) and I want to plot a heatmap to show their density distribution.First I plot the histogram by:

D=dlmread('PC1-PC2.txt');
X=D(:,1);
Y=D(:,2);
H=histogram2(X,Y,37);

and I get fig1 (attached) the h with the follow information:

Data: [5598x2 double]
           Values: [37x37 double]
          NumBins: [37 37]
        XBinEdges: [1x38 double]
        YBinEdges: [1x38 double]
         BinWidth: [0.1310 0.1390]
    Normalization: 'count'
        FaceColor: 'auto'
        EdgeColor: [0.1500 0.1500 0.1500]
    The range of h.XbinEdges is -2.3 to 2.5; and the range of h.YbinEdges is -2.8 to 2.3

and a 37 x 37 matrix counts by:

counts = h.Values

After that I applied every value in the matrix "counts" to an equation, and got my target 37 by 37 matrix "G" (attached), and then I plot the heatmap

imshow(heatMap, []);
axis on;
colormap(hot(256));
colorbar;

and I got fig2 (attached)

In order to make the image larger and smoother, I used the following codes:

heatMap = imresize(G, [700, 700])
imshow(heatMap, []);
axis on;
colormap(hot(256));
colorbar;

Now the image was close to what I want, but the X and y axes ranges (0 to 7000) are meanless. I hope the axes ranges can reflect the ranges of the original coordinate data, just like the histogram （fig1): x axis range -2.3 to 2.5 and y axis range -2.8 to 2.3. So I passed these range to imshow:

 imshow(heatMap, 'XData',[-2.3 2.5],'YData',[-2.8,2.3]);
 axis on;
 colorbar;
 colormap(hot(256));

and I got fig4 (attached)

Obviously the image is wrong. Is it possible to obtain a image like fig3 but have correct axes ranges like fig4?

Answer 1

[John BG]

Yeping

the range in imshow( , ,[range]) does not change the axis ticking.

For such purpose, to change the X axis ticks to [-2.5 2.3] and Y axis to [-2.8 2.3] you have to capture the axis handle, and access the right fields called XTickLabel and YTickLabel.

First.-

modify your code, where you imshow the heatMap adding the handle or pointer f3:

    f3=figure(3);imshow(heatMap, []); axis on; colormap(hot(256)); colorbar;

now capture the handle to the axis of the figure 3:

ax3=f3.CurrentAxes

have a look to the current X ticking:

Lx=ax3.XTickLabel
Lx = 
    '100'
    '200'
    '300'
    '400'
    '500'
    '600'
    '700'

You say you want [-2.3 2.5] range, let's have the X axis ticking with same amount as it is for [100 700], with numel(Lx)

Lx2=linspace(-2.3,2.5,numel(Lx))

Because cells have variable length, use this loop to have '-' and other length variations correctly converted to cell type:

Lx_cell={};
for k=1:1:numel(Lx)
    L1=num2str(Lx2(k));
    Lx_cell=[Lx_cell L1];
end

Lx_cell=Lx_cell';

Now directly apply the new tick labels

f3.CurrentAxes.XTickLabel=Lx_cell

f3 = 
  Figure (3) with properties:

        Number: 3.00
          Name: ''
         Color: [0.94 0.94 0.94]
      Position: [501.00 161.00 872.00 795.00]
         Units: 'pixels'

the command window doesn't show, but now

f3.CurrentAxes.XTickLabel
ans = 
    '-2.3'
    '-1.5'
    '-0.7'
    '0.1'
    '0.9'
    '1.7'
    '2.5'

check your figure

And for the Y axis

Ly2=linspace(-2.8,2.3,7)

Ly_cell={};
for k=1:1:7
    L2=num2str(Ly2(k));
    Ly_cell=[Ly_cell ;L2];
end

f3.CurrentAxes.YTickLabel=Ly_cell

Yeping

If you find my answer useful would you please be so kind to mark it as Accepted Answer?

To any other reader, please if you find this answer of any help solving your question,

please click on the thumbs-up vote link,

thanks in advance

John BG

<mailto:[e-mail] [e-mail]>

[Image Analyst]

The gray level range doesn't change the tick marks and labels, but the XData and YData properties do. So that makes it real easy - you don't need to fool with tick labels manually.

imshow(heatMap, [], 'XData', h.XBinEdges, 'YData', h.YBinEdges);

Answer 2

[Image Analyst]

Try adding

caxis([0, 13]);

Answer 3

[Image Analyst]

Yeping, this seems to work just fine.

D=dlmread('PC1-PC2.txt');
X=D(:,1);
Y=D(:,2);
h = histogram2(X,Y,37)

counts = h.Values;
G = counts/12; % Or whatever you did.

heatMap = imresize(G, [700, 700]);
imshow(heatMap, [], 'XData', h.XBinEdges, 'YData', h.YBinEdges);
axis on;
colormap(hot(256));
colorbar;

Is there anything wrong with it?

[Yeping Sun]

By this code I get the attached image. Now the axes are correct. But I think your color gradient is more good-looking than mine. how do you get this color effect?

[Image Analyst]

I used hot(256) to create the colormap - same as what it looks like you did. It looks like your data is different though.