# calculating hourly means of a matrix

2 views (last 30 days)
Richard on 26 Aug 2012
Commented: Venkata on 11 May 2017
I have a dataset in the following format:
dd = datenum('2009-04-17 02:00');
ddend = datenum('2009-11-24 12:27');
Day = dd:(1./24./60):ddend;
time = datevec(Day);
data = rand(length(time),12);
Where time refers to the time (in minutely intervals) and data refers to a matrix of data points. I would like to find the best possible way of calculating the hourly means of each column of the matrix (i.e. mean of every 24 rows in each individual column). How would I achieve this, by taking into consideration the time variable?

Jan on 27 Aug 2012
limit = floor(length(time) / 60) * 60;
data_hour = reshape(data(1:limit), 60, []);
data_mean = mean(data_hour, 1);

Andrei Bobrov on 27 Aug 2012
other variant
[c,c,c] = unique(time(:,1:4),'rows');
s{2} = kron((1:size(data,2))',ones(size(data,1),1));
s{1} = kron(ones(size(data,2),1),c);
result = accumarray(s,data(:),[],@mean);
Venkata on 11 May 2017
Dear Andrei Bobrov,
Could you please elaborate on your code. I want to reshape hourly mean data which is in vector form to a matrix, but there might be some data (hourly values) missing. When I tried your code, I got an error related to array size. Mine is a large array with 210888 rows.
Thank you.