Time stamp for discrete Fourier transform

Hi all,

I am trying to create a time stamp for my weather data. I took my readings every three hours for simplicity sake lets say I just want four days and a half of August how can I do this? Attach is my attempt with my previous try.

Thanks in advance

a=[ 27 26 24 29 35 38 37 35 31 28 27 31 36 39 39 35 35 32 28 26 30 37 40 39 36 33 ]
a1=mean(a)%mean of matrix A
a2=a-a1;%subtraction of the mean at each observation
t1=datetime(2015,8,2);%time stamp attemp 1
t2=datetime(2015,8,6);
t1:hours(3):t2
y=fft(a2);

Answer 1

[Star Strider]

Try this:

a=[ 27 26 24 29 35 38 37 35 31 28 27 31 36 39 39 35 35 32 28 26 30 37 40 39 36 33 ]
a1=mean(a)%mean of matrix A
a2=a-a1;%subtraction of the mean at each observation
t1=datetime(2015,8,2);%time stamp attemp 1
t2=datetime(2015,8,6);
tv = t1:hours(3):t2;
L = numel(a);                                                   % Nr Samples
Ts = 3/24;                                                      % Sampling Interval (Units: Days)
Fs = 1/Ts;                                                      % Sampling Frequency (Units: Cycles/Day)
Fn = Fs/2;                                                      % Nyquist Frequency (Units: Cycles/Day)
y=fft(a2)/L;
Fv = linspace(0, 1, fix(L/2)+1)*Fn;                             % Frequency Vector
Iv = 1:numel(Fv);                                               % Index Vector

figure
plot(Fv, abs(y(Iv))*2)
grid
xlabel('Frequency (Cycles/Day)')
ylabel('Amplitude')

Experiment to get different results.

[desert_scientist90]

Thanks! I was kinda lost with this but now I understand!

[Star Strider]

As always, my pleasure!