Question

I have 3 variables (x, y, z) each in a 192×1 vector. The data is all random and sometimes has missing values (NaNs).
I would like to plot variable c against the other two variables x and y as contour plot (x on the x-axis, y on the y-axis, and z making up the contour lines).

My biggest problem is making up the 192×192 matrix for x and y because x and y consist of random data. I have tried to use GRIDDATA, Delaunay, and TriScatteredInterp but they did not work because I have NaNs in my data.
Does anyone know what I can do? If there was a way that I could return the matrix of x vs y from the scatter plot of x and y then my problem would be solved.

Example code

x=rand(192,1);
y=rand(192,1);
z=rand(192,1);

[X,Y]=meshgrid(x,y);
contour(X,Y,z);

??? Error using ==> contour at 74
Z must be size 2×2 or greater.

Answer 1

What are you going to do with NaNs? Do they contain any additional information to draw the contour plot? Cannot you just remove them?

idx = any(isnan([x y z]),2);
x(idx)=[];
y(idx)=[];
z(idx)=[];

Then you can apply solution from this question. The missing data will be interpreted anyway.