One dimensional interpolation of a two dimensional matrix

I have a matrix of 24x365, i want to increase it to 1440x365 by interpolating 59 values in between every two values rowwise. because i have a hourly data of one year. (24 hours) 365 (days). I need minute wise data so one point has to become 60 points..

[Stephen23]

The expected output size does not match your explanation. If you add 59 values between any two existing adjacent values (exactly as your state), then you will actually get an output matrix with

23*60+1 % rows

ans = 1381

not

24*60 % rows (incorrect)

ans = 1440

This is probably because you have not considered the ends (e.g. requiring extrapolation before/after the existing values).

"I need minute wise data so one point has to become 60 points."

Have you considered what happens at the edges of your data?

Here is a simple demonstration of your explanation and the corresponding output matrix size. Here is a 1x4 vector:

X = [0,    3,    6,    9];

Now lets add two points between adjacent existing values, exactly as you specify:

Y = [0,1,2,3,4,5,6,7,8,9];

The output vector actually has

N = numel(X); % number of existing points
P = 2;        % number of points added
(N-1)*(P+1)+1 % number of output points

ans = 10

Lets check this value against your estimation:

numel(Y)

ans = 10

N*(P+1) % your approach

ans = 12

[Mohammed Yousuf]

Your explanation is appreciated and thanks for it. Now I understood.

Thank you very much

Answer 1

[KSSV]

Let A be your 24*365 data matrix.

t = (60:60:24*60)' ; 
ti = (1:1:24*60)' ;
B = zeros(1440,365) ;
for i = 1:365
    B(:,i) = interp1(t,A(:,i),ti) ; 
end

[Mohammed Yousuf]

Mostly it works but not completely because. First 60 rows are NaNs. How to avoid it. First value is statrting at 61st row.

[KSSV]

Either the first sixty or last sixty have to be extrapolated....we cannot gurantee that value.

[Mohammed Yousuf]

Thanks for the answer

I got it