3D plot matlab from X Y Z data and integrating a slice function

Hi I have a data formated in CSV file (unfortunately I cannot upload it cause it exceeds 5MB even if I compressed the file). However, my data is composed of 3 coordinates system XYZ which is assigned to a certain energy.

e.g.

X Y Z Energy
0  0 1   48

having a dimension of 100 x 100 x 100

I tried this following code

###

Array=csvread('processedata.csv');
X = Array(:, 1);
Y = Array(:, 2);
Z = Array(:, 3);
data = Array(:, 4);
colormap turbo 
s=scatter3(X,Y,Z,20,data,'filled','s','MarkerEdgeColor','none')    % draw the scatter plot
alpha(s,.1) 
ax = gca;
ax.XDir = 'reverse';
%view(-31,14)
% create and label the colorbar
cb = colorbar;
caxis([0 38.1])
xlabel('X')
ylabel('Y')
zlabel('Z')
cb.Label.String = 'Dose';

###

and the output is this

but I wanted to show like this instead with a grid box and slices so that could see what is happeninng inside the box

Answer 1

[Walter Roberson]

If you have a regular grid of x y z points, then pull out the columns and reshape() and possibly permute() into an appropriate sized 3D grid of energy.

If the points are not regular, then pull out min and max x y z and linspace() vectors and ndgrid() to construct matrices X Y Z of coordinates to approximate the data at. Then scatteredInterpolant(x, y, z, Energy) and evaluate the resulting function at X Y Z to get a grid of energy.

Either way, you now have a 3D grid of energy that you can use slice() with.

[Denxybel Montinola]

I have attached the file.

I have followed your innstruction but seems the slice needs to have an input of a volume

Array=csvread('processedata.csv');
X = Array(:, 1);
Y = Array(:, 2);
Z = Array(:, 3);
D = Array(:, 4);

A = reshape([X;Y;Z],[],3);
I=scatteredInterpolant(X1,Y1, Z1, D)
slice(X1,Y1,Z1, A, 0, [],[])

[Walter Roberson]

You have 50 unique X values, 100 unique Y values, 100 unique Z values. They form a grid of values. However, instead of 500000 points, you have over 900000 points -- so you have about 1.8 values on average for each location. scatteredInterpolant will average the values at any one location.

Array=csvread('processedata.csv');
X = Array(:, 1);
Y = Array(:, 2);
Z = Array(:, 3);
D = Array(:, 4);

ux = unique(X);
uy = unique(Y);
uz = unique(Z);
[XG, YG, ZG] = meshgrid(ux, uy, uz);
F = scatteredInterpolant(X, Y, Z, D);
DG = F(XG, YG, ZG);

slice(XG, YG, ZG, DG, 0, [], [])

[Walter Roberson]

You have a lot of points, about half a million. Trying to plot them all would be really slow. The range is about -2800 to +89 . Even just an isosurface at +80 is slow as there are over 10000 irregularly-placed points above +80. It is difficult to see any structure in the data. I recommend that you

histogram(DG(:), 100)

and look at the distribution (and keep in mind that only 1% population would still be ~5000 points)

[Denxybel Montinola]

I am actually trying out different representations of my data and I agree about your point that I can hardly see any structure of my data. However, I have see some pattern from it. A big thanks for your help Walter, you're the best!

[Walter Roberson]

I think you might need to project your 4D data (three independent, one dependent) into 3D so that you have a chance of seeing continuous changes in a recognizable way. isosurfaces look just too irregular to make sense out of.