Plotting two vectors for a period of time

Hi,

I have a very simple question. I have two vectors x and y, which I plot against each other. I have another vector P which shows the evolution of the values of x against time. I want to extend the x vs y plot for the time given and highlight the values of P at each time. This should look just like the extension of my original curve, but with a line running across on it through time.

[Osama Maqbool]

As an example, x=[1.1 2.4 5 6 9.6] y=[0.2 0.3 0.5 0.67 0.9] this is the data for a characteristic curve of a device. Now I have the values x takes for 100 seconds and I want to show x,y and t on a 3d plot. And on every t, I want to point the value x has and the corresponding value of y.

Answer 1

[Robert]

% From your question:
N = 100;
x = [1.1 2.4 5.0 6.00 9.6];
y = [0.2 0.3 0.5 0.67 0.9];

% A little more setup
time = (1:N)/4;
xy_index = randi(length(x),size(time)); % The active (x,y) index at each time

% To make line crawl across surface instead of through it, we need to do
% some interpolation. We will plot the original data with markers to
% hilight it and plot the interpolated data with lines.

% The index for x and y might move from any index to any other index. Here
% we use the diff to see how far it moved and fill in the spaces in
% between. So a jump from 1 to 4 is replaced with the indices [1,2,3,4]. If
% there is no jump (2 to 2 for example) we still need a spot for the second
% value, which is why we use max(1,abs(a)) below.
% 
% Because each pair of indices will be replaces with an array of unknown
% length, we use arrayfun to package all the ragged arrays in a cell array,
% then concatenate them together in the second line.
xy_index_interpd = arrayfun(@(a)sign(a)*ones(1,max(1,abs(a))),diff(xy_index),'Uniform',false);
xy_index_interpd = [xy_index_interpd{:}];
xy_index_interpd = cumsum([xy_index(1),xy_index_interpd]);
% This is very much like the above lines, except that we are calculating how
% much we need to advance the time for each partial step across the
% surface. For this we use 1/number of partial steps so that the partial
% steps always add up to a whole step in the end.
t_index_interpd = arrayfun(@(a)ones(1,max(1,abs(a)))/max(1,abs(a)),diff(xy_index),'Uniform',false);
t_index_interpd = [t_index_interpd{:}];
t_index_interpd = cumsum([1,t_index_interpd]);
% Then we get our time values by interpolating time against the new
% indices.
t_interpd = interp1(1:N,time,t_index_interpd);

% And finally, we plot the results. A surf with x,y extended over time and
% a line to hilight the values of (x,y) over time.
surf(x,time,repmat(y,[length(time),1]),'EdgeColor','none') % y is expanded for surf
hold on
% The easy part
plot3(x(xy_index),time,y(xy_index),'ro')
% The harder part (for which we did all that interpolation above.
plot3(x(xy_index_interpd),t_interpd,y(xy_index_interpd)+max(abs(y))/1000,'r-') % ever so slightly above the surface for cleaner rendering
hold off

[Robert]

In case you are wondering why we would do all that work, here is the same plot using only

plot3(x(xy_index),time,y(xy_index),'r-o')

as the trace across the surface.

[Osama Maqbool]

That did it. Thanks a lot!