Irregular time series data interpolation

Hi all,

I have a time series data: (date, time, tide height)

11/4/2013  2:49  0.39
11/4/2013  8:59 -0.39
11/4/2013  15:12 0.4

etc.

How to interpolate the data so I can get tide height values with interval of 10 minutes?

Thanks Andi

Answer 1

[Andrei Bobrov]

f = fopen('yourdata.txt');
c = textscan(f,'%s %s %f');
fclose(f);

DT = datenum(strcat(c{1:2}),'mm/dd/yyyyHH:MM');
dt = 1/24/6; % (10 minites)/day' 
p = ceil(DT(1)/dt);
[y,m,d,h,mm] = datevec(p*dt);
dn = datenum(y,m,d,h,mm+10*(0:floor(DT(end)/dt) - p)');
dv = datevec(dn); 
out = [dv(:,1:5),interp1(DT,c{3},dn)];

[Andiyu]

Hi Andrei,

Thank you for the answer..but I still got error

Error using datenum (line 179) DATENUM failed.

Caused by: Error using datenum (line 165) Incorrect number of arguments

Can you please explain?

Cheers, Andi

[Andiyu]

solved the problem

f = fopen('data.txt');

c = textscan(f,'%s %s %f');

fclose(f);

DT = datenum(strcat(c{1:2}),'dd/mm/yyyyHH:MM');

dt = 1/24/6; % (10 minites)/day'

p = ceil(DT(1)/dt);

[y,m,d,h,mm,ss] = datevec(p*dt);

dn = datenum(y,m,d,h,mm+10*(0:floor(DT(end)/dt) - p)',ss);

dv = datevec(dn);

out = [dv(:,1:5),interp1(DT,c{3},dn)];

[ceb]

Hi Andrei and or Andiyu, It's an older thread but if you are still around... I was looking to do basically the same thing as Andiyu except interpolate 1 minute data from 10 minute data. I was able to adjust your above code to get (almost) what I am after but changing the following line: dn = datenum(y,m,d,h,mm+10*(0:floor(DT(end)/dt) - p)',ss); to dn = datenum(y,m,d,h,mm+1*(0:floor(DT(end)/dt) - p)',ss);

BUT the output is only as many rows as the original data - meaning I am still missing interpolated data. Original data is from 5/2/2015 to 26/2/2015 but results only go from 5/2/2015 to 7/2/2015. I am not sure what else in the code needs to be changed... Any suggestions? Thank you!