pointsize = 15;
scatter(x(:), y(:), scalars(:), pointsize, scalars(:))
However, if what you started with was a rectangular grid of data, and you rotated the axis, so what your data is "really" is still rectangular, then you should consider using mesh() or pcolor().
Supposing your original rectangular grid is orig_x and orig_y, then
[OX, OY] = ndgrid(orig_x, orig_y);
[OXYv] = [OX(:), OY(:)];
now do the rotation on the pairs x=OXYv(K,1), y=OXYv(K,2) to get Rxv, Ryv, vectors of the rotated values, in the same order as OXYv. Then
Rx = reshape(Rxv, size(OX));
Ry = reshape(Ryv, size(OX));
and then you can
pcolor(Rx, Ry, scalars)
Or you could have just taken your rectangular image, done an imrotate() on it to bring it into alignment with the axis, and image() the result.
Third method: make a hgtransform(), image() the rectangular matrix with 'Parent' being the hgtransform, then set the transform matrix to display the rectangular image at a tilt to the axis.
