To take mean/average array values xyz(4,:) or xyz(5,:) at xyz points to reduce number of data points/averaged

I am interested to reduce the data array as big as xyZ=9 X 275416 size, with first three column having xyz co-ordinates. let take mean values at each near by x, y , z values and make size small. The mean will be for xyz(4,:), xyz(5,:) and so on which are other than co-ordinate . This will make it easy to analyse in scatter plot(as when I plot now one color overlap other)

I try using the following : where z1 is = xyz(3,:) z -cordinate points , y1 is y co-ordinates, y1 and z1 used to get the index for the many points at near by for each grid x y z values in reduced scale. z and xo are the reduced grid points.

But its too much for big data time is slow

% vector array to less number of points
z= 0.00:0.01:0.1; % divide z and X to some number of grid points for 3D space
xo=-0.05:0.0004:0.05;
id =[]
k=1;v2=0
for i=1:length(z)
    for j=1:length(xo)
idx=find(z1+z(i)>-0.0005 & (z1+z(i)<0.0005) & y1-xo(j)>-0.0005 & y1-xo(j)<.0005) % dz +dx   % index location nearest to point array to xyz2
if (isempty(idx)==1)
    v2(k)=0;
w2(k)=0;
y2(k)=xo(j)%mean(y1(idx)); %
z2(k)=z(i) %mean(z1(idx));%
else
y2(k)=mean(y1(idx)); %
z2(k)=mean(z1(idx));%
v2(k)=mean(v(idx));
w2(k)=mean(w(idx));
end
k=k+1;
end
end

I wannt something which is fast : can we use vector as condition for find() function, as that will return index - idx as vector/1D array then I can remove the for loop and make it in time. The mean also can be done as mean vector