How to obtain a histogram for 3d data?

Hello!

I'm curious about an extension of the hist3 function in matlab. For a data set in 2D, hist3 is a 3D representation shaped like the data. For example, for matrix A, I get the results:

load A
plot(A(:,1),A(:,2),'linewidth',2); axis tight; grid on;
figure; hist3(A,'Nbins',[30 30],'CDataMode','auto','FaceColor','interp');

But if we consider a 3D data input like B:

load B
plot3(B(:,1),B(:,2),B(:,3),'linewidth',2); axis tight; grid on;

Is there any function that can give a representation shaped like the data based on histogram? Something similar to hist3? To show the histogram and the shape of the data too?

Thank you!

Answer 1

[Torsten]

Since your curves are all 1d, a 1d- histogram suffices if you identify x-axis coordinate with curve length or parametrized points in 2d resp. 3d.

A histogram for 3d-data in general is not possible. It would be a representation in 4d - and humans live in 3d.

[Jessie Bessel]

Thank you for your answer. Yes, I know that a clasic histogram for 3d is not possible. I am looking for a histogram based method to highlight the contour/shape/boundary of a 3d representation.

[Torsten]

Yes, that's why I suggested a 1d-histogram. For a presentation, show the curve in 3d where the distribution lives together with the 1d-histogram in which the x-axis represents curve length.

[Jessie Bessel]

Can you help me with a little exemple based on my data, please? I didn't quite understand the processing part.