how to do density plot?

suppose, I have a text file 'data.txt' that has three columns and contains output of below.

    % code
fileID = fopen('data.txt','w');
for x = 0.1:0.1:10
     for y = 0.1:0.1:10
           z = x*y;
           T = [x, y, z];
           fprintf(fileID,'%f %f %f\n',T);
     end
end
  fclose(fileID);

Now i want to do the density plot of this data. Please help.

[Adam Danz]

This code writes data to your data.txt file. Option 1 is to read the file to get the data back into matlab and then do the plotting. Option 2 is to save the data within this function/script so you don't have to read it back in. Depending on what you're trying to do, you may not even need to write it to the txt file in the first place (option 3).

Secondly, you need to specify what data you want to plot. For example, do you want to plot the density of each column of T separately?

Thirdly, how would you like to represent the density? Are you looking to plot a histogram?

[sourabh mittal]

I want option 1. I want 1st column of T on x-axis and 2nd column on y-axis and then 2-D color density plot of 3rd column with a color bar.

[Adam Danz]

I continued in the answer section though I'm unsure about the plot you're describing.

[sourabh mittal]

This is what i was searching for

   % code
fileID = fopen('data.txt','r');
formatSpec = '%f';
A = fscanf(fileID,formatSpec);
x=A(1:3:end); y=A(2:3:end); z=A(3:3:end);
f=scatteredInterpolant(x,y,z);
[X,Y]=meshgrid(0.1:0.1:10,0.1:0.1:10);
Z=f(X,Y);
surf(X,Y,Z)
view(0,90)
colorbar

Thank you!

Answer 1

[Image Analyst]

Try this vectorized approach with meshgrid():

% Generate the data.
xv = 0.1:0.1:10
yv = 0.1:0.1:10
[x, y] = meshgrid(xv, yv)
z = x .* y
T = [x(:), y(:), z(:)]

% Show the x,y,z data as a surface
surf(xv, yv, z, 'EdgeColor', 'none');
xlabel('x', 'FontSize', 20);
ylabel('y', 'FontSize', 20);
zlabel('z', 'FontSize', 20);

% Write the file.
filename = fullfile(pwd, 'data.txt');
fileID = fopen(filename, 'wt');
fprintf(fileID,'%f %f %f\n',T'); % Don't forget to transpose with '
fclose(fileID);
type(filename);
% delete(filename);

[sourabh mittal]

This is nice approach but I want to read the file to get the data back into matlab and then do the 2-D plotting.

[Image Analyst]

OK. I thought you were going to visualize the data in the same program that created it. But it looks like Adam alerted you to the function dlmread(). But you might still look at my vectorized way of using meshgrid() - it's a handy function that allows you to do in one line what would otherwise be a multi line set of nested for loops.

Anyway, after calling dlmread(), you can use surf() like I showed, or scatter() or imagesc() like Adam showed, or even histogram() or histogram2(), depending on what you mean by density. Still even now, no one has a clear idea of what you mean by that, other than your acceptances of Adam's answer which I guess means you wanted either scatter() or imagesc().

[sourabh mittal]

Thanks my problem was solved by rotating surf by 90 degree.

Answer 2

[Adam Danz]

You can read the data back in with dmlread()

data = dlmread('data.txt', '%f');

' data' will be the [10000 x 3] matrix.

For the figure, I'm not sure if this is what you're looking for but you can try it.

figure; 
imagesc(data(:,1), data(:,2), data(:,3)); 
colorbar

Another option

figure; 
scatter(data(:,1), data(:,2), 10, data(:,3))

[Adam Danz]

See comments under the answer provided by @ImageAnalyst for more interpretations of what you'd like to plot in 2D space.