Scatterplot equivalent in plotting 3d data in 2d

Hi everyone,

I have three vectors of same size (54796X1 double) and want to plot them like the image attached. Using the scatter function, scatter(x,y ,35, z, 'filled'), naturally I get a plot made of dots. I increased the dot size to make the data look full but as you can see it is not a good solution. Do you know how the plot in the attached image is generated with the same data? I thought of image but I can not get it to work. I have attached the vectors in the workspace. Here is how I got my plot using scatter and contour:

     scatter(x,y ,35,  z, 'filled') 
     x_interpolated = linspace( -90, 90, 300 );
     y_interpolated = linspace( -90, 90, 300 );
     [ hx, hy  ] = meshgrid( x_interpolated, y_interpolated );
     F = TriScatteredInterp(x, y, z);
     data_interpolated = F(hx, hy);
     hold on
     contour( x_interpolated, y_interpolated, data_interpolated, 10, 'k' );

Any help is highly appreciated.

/Hasti

[KSSV]

Attach the data.....so that we can work on it and give a solution.

[Walter Roberson]

Those appear to be contour plots not scatter plots.

[Walter Roberson]

Scatter is not relevant here. You do not have hold on, so your entire plot is replaced by the contour plot derived from the scattered Interpolant

[Hasti Yavari]

Hi walter, Thanks for fixing the code section in my question. I do have hold on and if you run scatter and contour separately you will see that they are two different plots.

[Walter Roberson]

It looks to me to be likely that your scatter plot is completely hidden by the contour plot. Both are in the z=0 plane and since they are the same depth and are both space filling areas instead of lines, the one that is going to show up is the one drawn last.

[Hasti Yavari]

Hi Walter, I understand your concern but that is not the case here. I can not upload anymore images due to mathworks regulations but if you just run the contour and scatter function lines separately you will see my point.

Answer 1

[KSSV]

You need to do interpolation with the scattered data...and then use surf or pcolor. Read about scatteredInterpolant

[Hasti Yavari]

When I try your suggestion, I just get a mostly black plot. At which points do you calculated the z_interpolated? Here is my attempt:

F =scatteredInterpolant(phi_part, theta_part, data);
   x_interpolated = linspace( -90, 90, 300 );
   y_interpolated = linspace( -90, 90, 300 );
   [hx, hy] = meshgrid( x_interpolated, y_interpolated );
   data_interpolated = F(hx, hy);
   surf(data_interpolated) 
   view([ 90 90 ]);

[Walter Roberson]

Add 'EdgeColor', 'none' to the surf call

[Hasti Yavari]

Turning off the edge color takes away the black but it still does not give the plot I am looking for.

[KSSV]

Try shading interp after surf

[KSSV]

I already mentioned to attach the data....

[KSSV]

load dataforplotting3din2d.mat ;

m = 500 ; n = 500 ;
x_interpolated = linspace( min(x), max(x),m);
y_interpolated = linspace(min(y),max(y),n);  [ hx, hy  ] = meshgrid( x_interpolated, y_interpolated );
F = TriScatteredInterp(x, y, z);
data_interpolated = F(hx, hy);
hold on
pcolor( x_interpolated, y_interpolated, data_interpolated);
shading interp 
contour( x_interpolated, y_interpolated, data_interpolated, 30, 'k' );

[Hasti Yavari]

I see! Thanks KSSV. So the solution was right in front me but I just didn't see it><

[KSSV]

Yes....you were taking a small region out of the given region.

[Walter Roberson]

... and no scatter() plot.