Interpolate to plot contour in plane of non-uniform 3D data

I have a set of three dimensional data (attached) that is on a non-uniform grid. In the code below, you see how I load, reshape, and plot to get an accurate 3D representation.

clear; close all

data = xlsread('C:\PATH\V27TestOutputCelementCV27_WakeElemData_00206.csv');
node = data(:,2);
orgn = data(:,3);
xr = data(:,4);
yr = data(:,5);
zr = data(:,6);
uu = data(:,7);
vu = data(:,8);
wu = data(:,9);

m = length(orgn(orgn==1)); % the input file is sorted by origin node, so use it to reshape
n = orgn(end);

x = reshape(xr,[m,n]); % each column is a different origin node now
y = reshape(yr,[m,n]);
z = reshape(zr,[m,n]);
u = reshape(uu,[m,n]);
v = reshape(vu,[m,n]);
w = reshape(wu,[m,n]);

quiver3(x,y,z,u,v,w), xlabel('x/R'), ylabel('y/R'), zlabel('z/R')

I've looked around, but I can't figure out how I can create a grid, do some interpolating, and plot a contour slice of this data. For example, how would I plot a contour showing the u data along y = 0 for all x and z?

UPDATE

I got some of it with this:

F = scatteredInterpolant(xr,yr,zr,uu);
[xq,yq,zq] = meshgrid(-1:0.05:8,-1.5:0.5:1.5,-1.5:0.05:1.5);
uq = F(xq,yq,zq);

figure, contourf(squeeze(uq(4,:,:))'), colorbar

The data is new to me, but I think this looks "right". One thing that I'm not getting right is using x and y in contourf to get the axes scaled correctly. I tried contourf(squeeze(xq(4,:,:))',squeeze(yq(4,:,:))',squeeze(uq(4,:,:))'), but this produced a line. I'm missing something... Also, if there are better ways, I'm still interested.

[KSSV]

Read about slice.

[Daniel]

Slice is helpful, but a couple questions:

To get a 3D array of V, or in my case, u, I already interpolated it, didn't I? I guess I'm not clear on what scatteredInterpolant did, but I want to be sure I'm not interpolating twice since slice also interpolates. Second, is there a way to turn off the grid lines in the slice planes?

[Daniel]

For grid lines: https://www.mathworks.com/matlabcentral/answers/51167-slice-hiding-the-edges-of-slice

Answer 1

[darova]

• Use griddata to interpolate your data (create 3d matrices)
• Use slice to create crossection views

x = 2*rand(5000,1)-1; 
y = 2*rand(5000,1)-1; 
z = 2*rand(5000,1)-1;
v = x.^2 + y.^2 + z.^2;
[xq,yq,zq] = meshgrid(-1:0.1:1);
vq = griddata(x,y,z,v,xq,yq,zq);
plot3(x,y,z,'.r')
slice(xq,yq,zq,vq,[-1 1]/2,[-1 1]/2 ,[])
axis vis3d

[Daniel]

v is already part of the file I attached, so I can't just compute it from the outputs of meshgrid.

[darova]

Use griddata to calculate 3D matrices

[Daniel]

I think griddata is the best solution so far. If you'd like to edit your "answer", I can mark it as correct.