How to shift data sets automatically

I have two sets of data that produce curves (see attached), and I want to analsye them both so need the data points to line up. Is there a code I can use to get Matlab to line up the data points for me? i.e the peak point for U should be the same data point as the one for D. The matrices are the same size but can be shortened if necessary to do this.

[Vilém Frynta]

hi,

my approach to this problem would be:

• using max and find, find the highest value (peak) and it's position (index)
• using simple logic, find which maximum is "further" (which index is higher). then, find the difference between the distances of peak (difference in index numbers)
• using NaN, create vector full of NaNs, the length will be length of the original data + difference of the maximum indexes
• using indexing and knowledge about position of the peaks, insert your data into your NaN vector.

definitely would be doable and easier to show with code, if you attach your data. but, anyway, i hope my guidance was helpful to you.

[Damaris Litton]

Thank you @Vilém Frynta! I have attached my data, would it be possible for you to explain using the data sets? Thank you.

[Vilém Frynta]

hello,

yes.

% load data
load U.mat
load D.mat

% make vectors from your data (previously columns)
U = U'

U = 1×2328

0.0043 0.0043 0.0045 0.0044 0.0047 0.0046 0.0047 0.0046 0.0051 0.0049 0.0050 0.0047 0.0048 0.0049 0.0048 0.0047 0.0046 0.0044 0.0043 0.0044 0.0049 0.0048 0.0046 0.0045 0.0047 0.0049 0.0046 0.0047 0.0046 0.0046

D = D'

D = 1×2328

0.0139 0.0139 0.0139 0.0139 0.0139 0.0139 0.0139 0.0139 0.0139 0.0139 0.0139 0.0139 0.0139 0.0139 0.0139 0.0139 0.0139 0.0139 0.0139 0.0139 0.0139 0.0139 0.0139 0.0139 0.0139 0.0139 0.0139 0.0139 0.0139 0.0140

% find maxs (peaks)
idx.U = find(U == max(U));
idx.D = find(D == max(D));

% difference between the peaks
idx.diff = idx.U - idx.D

idx = struct with fields:
       U: 656
       D: 398
    diff: 258

% let's assume U is further
is_U_further = 1;

% find if D is further (if idx.diff is negative)
if idx.diff < 0
    is_U_further = 0;
    idx.diff = idx.D - idx.U % switch the difference in that case
end

% create NaN vector and new variables
L = length(U);
new.U = NaN([1 L+idx.diff]);
new.D = NaN([1 L+idx.diff]);

% fit the values so the peaks are at the same index
if is_U_further == 1; % if U is further... do this:
    new.U(1:L) = U;    
    new.D(idx.diff:idx.diff+L-1) = D; % move D further, so it matches
end

if is_U_further == 0;% if D is further...
    new.D(1:L) = D;
    new.U(idx.diff:idx.diff+L-1) = U;
end

plot(new.U)
hold on
plot(new.D)

thought i did it wrong but it looks like "D" is just too small (no pun intended). you can try to strech that or make new, relative Y axis for it.

Answer 1

[Mathieu NOE]

hello

myabe this ?

I took the options to have both peaks overlaid and also the two curves have same initial y value

This is the result

of course you can change the code if your prefer to have same y final value but I doubt you can have start, end and peaks points all 3 overlaid

% load data
load U.mat
load D.mat

% find maxs (peaks)
[umax,iumax] = max(U);
[umin,iumin] = min(U);
% dU = umax- umin;
dU = umax- U(1);

[dmax,idmax] = max(D);
[dmin,idmin] = min(D);
% dD = dmax- dmin;
dD = dmax- D(1);

% D  y dir stretch factor
D_ySt = dU/dD;
D = D*D_ySt;
% apply y shift to have both curves start at same value U(1)
yshift  = D(1) - U(1);
D = D - yshift;

% x axis shift to align both peaks (delay D vs U)
xshift = iumax - idmax;

% define x axis vectors 
xu = 1:numel(U)
xd = xu + xshift;
plot(xu,U,xd,D)