3D surface plot from imported 2D excel array

New to Matlab. Im sure this is answered somewhere already, but I can't manage to find it; I have a 2D array of experimental data tabulated in excel. 1st row and column are varied parameters, and I measure the corresponding amplitudes in the matrix. Importing the table to Matlab is simple. I would then like to plot a 3D surface of the 1st row (x) vs. 1st column (y), for the amplitudes in the matrix. This is very straightforward in excel, but the plots look terrible. Thanks in advance for the help.

Answer 1

[Star Strider]

Try this:

[D,S,R] = xlsread('ExampleArray.xls');

figure(1)
ribbon(D(:,1), D(:,2:end))
grid on

[X,Y] = meshgrid((1:size(D,2)-1), D(:,1));
Z = D(:,2:end);

figure(2)
surf(X, Y, Z)
grid on

[joe pearson]

Hi Star Strider, Thanks for the help! This is pretty close to what I am after. However, I still have a couple issues with the result. The x axis looks correct (first column of the table), but the y axis is from 1 to 5, and not the values in the first row of the table. Also the colormap is not correct. I tried this, but the colormap scaling is apparently not being automatically adjusted;

[D,S,R] = xlsread('ExampleArray.xls');
[X,Y] = meshgrid((1:size(D,2)-1), D(:,1));
Z = D(:,2:end);
figure(2)
C=X.*Y;
surf(X,Y,Z)
colormap parula
shading interp
colorbar

[Star Strider]

As always, my pleasure.

The first value in the first row is NaN, and that will cause problems. That is the reason I chose [1:5] to test the code. The first column is also significantly different from the others, and since I have no idea what the data are, I used it for the independent variable.

You can get some color definition by adding:

caxis([0.001 0.025])

after the surf call. (No magickal knowledge to that — I experimented until I found something that worked!)

[joe pearson]

The caxis function worked nicely. It seems the amplitude of the Z data was too low for the colormap to rescale properly by default. I still haven't come up with a good way to rescale the x axis. It is actually the y axis that is being set correctly already (The 1st column of the data table). You are right they are the independent variable. The second independent variable, x, is the 1st row (wavelengths in this case). The Z amplitudes are the measured values in the rest of the table for the corresponding x,y settings. It seems like there should be a very straightforward way to deal with this, but I haven't managed to figured it out yet. Any help would be greatly appreciated.

[Star Strider]

The first element in the first row is NaN.

What should that wavelength be?

[joe pearson]

Yes, the way the data is generated in the excel file it is not supposed to be a number. The 1st row basically contains the column headings for the z data below, and the 1st column is the row headings. The y-axis looks correct (4000 to 6000 increments), but the 1-5 on the x axis should instead be the wavelengths 825-945 in the table.

[Star Strider]

I’m glad caxis works!

So perhaps:

Row_1 = linspace(825, 945, 5);

Would that do what you want?

[joe pearson]

That's almost what I wanted. I finally figured out how it works. Thanks again for all the help!! This is the code I am using, in case the next person to come along finds it useful;

[D,S,R] = xlsread('ExampleArray.xls');
Z = D(2:end,2:end);
M = max(max(Z));
figure(1)
X = D(1,2:end);
Y = D(2:end,1);
x1 = min(X);
x2 = max(X);
y1 = min(Y);
y2 = max(Y);
surf(X,Y,Z)
axis manual
xlim([x1 x2])
ylim([y1 y2])
colormap parula
shading interp
colorbar
caxis([0 M])
%view(2)

uncommenting the view(2) makes a nice psuedo 3D view

[Star Strider]

As always, my pleasure!