How do I calculate the mean value for the days in a big data set?

Hi I have a big data set 46933x12 and the data is collected every 5 minuets and the dates looks like

2014-01-01 00:00:00

2014-01-01 00:05:00

2014-01-01 00:10:00

... I want to have mean values for each of the 11 columes. How do i do this?

[John Chilleri]

You can acquire the mean of each column individually with,

for i = 1:12
   means(i) = mean(data(:,i));
end

Hope this helps!

[KL]

do you want to calculate the daily mean for these 11 variables?

[Johan Wilson]

Yes, so I hope to end up with a mean value for each day. so 162 mean values for every variable.

[KL]

if the timestamp is always uniform like from 00:00:00 to 23:55:00, the job is much easier, is that the case for you?

[Andrei Bobrov]

Please attach small example of your data file.

[Johan Wilson]

Here is a small sample of 60 data points.

Answer 1

[KL]

Here is a sample code with 3 variables and 5 days with uniform 5 minute timestep

a = rand(288*5,3);
t = (1:288:288*6)';
amean = zeros(5,3);
for i=1:length(t)-1
   amean(i,:) = mean(a(t(i):t(i+1)-1,:));
end

Answer 2

[Andrei Bobrov]

T = table2timetable(readtable('Test.Dataset.xls','ReadVariableNames',false));
out = retime(T,'daily','mean');

or for MATLAB <= R2016a

T = readtable('Test.Dataset.xls','ReadVariableNames',false);
[a,b,c] = datevec(T.Var1);
[dv,~,t] = unique([a,b,c],'rows');
[x,y] = ndgrid(t,1:size(T,2)-1);
out1 = [dv,accumarray([x(:),y(:)],reshape(T{:,2:end},[],1),[],@mean)];

[Johan Wilson]

I downloaded mathlab 2017 had 2015 before. It looks like it works for the small sample but it dont work for the full set. This is a slightly bigger one.

Answer 3

[Baptiste Ranguetat]

meanData = zeros(1,11);
for i=1:11
    meanData(i) = mean(data(:,i));
end