Convert 1 min average data to hourly average data

Hello!

I have data in the following format:

mm-dd-yy HH:MM m/s

There are some missing entries in this data, and I need to convert this to hourly average data. I have data for a whole year and doing it manually or just with for and if statements will be extremely time consuming. Could anyone please help me with this to get it done in the most efficient way?

Answer 1

[Walter Roberson]

If you have a new enough MATLAB, you can read the data using readtable() and convert the resulting table() object to a timetable() object which then has direct functions to group by time.

Otherwise, the easiest way to read the date and time depends on which MATLAB release you have. Also, is the file a .xls or .xlsx or a .csv ?

[Mehran Mustafa]

Walter,

I have MATLAB 2013a. And the file im using currently is .xlsx. I can convert it to .csv if its required.

[Walter Roberson]

Okay, could you do

[num, raw] = xlsread('YourFile.xlsx', 'basic');

and show the output

num(1:2,:)
raw(1:2,:)

and then repeat the same thing but without the 'basic' option, and show the result?

[Mehran Mustafa]

It did not work with 'basic' option. It said worksheet 'basic' not found. So i tried without 'basic' and got the following output for num(1:2,:) raw(1:2,:):

Also, now I have MATLAB 2016a as well. I have the function readtable() but not timetable().

[Walter Roberson]

I should have suggested

[num, raw] = xlsread('YourFile.xlsx', [], [], 'basic');

compared to the same thing without the 'basic' option, for R2013a.

Answer 2

[Chad Greene]

If you can get time into datenum format, downsample_ts makes it easy.

[Kanachi]

Hi,

I have the timestamp of my timeseries data for hydrophone data in datenum format. Each record is 8min duration. I would like to do 3hourly averages for my data and look at the plots. How do i got about. I have obtained the power spectrum levels using pwelch and got the corresponding frequecies too. Please advise. THANK YOU.

Answer 3

[Ohad Zivan]

if you have a timetable (called OldData here) in matlab it should be easy enough.

NewData=retime(OldData ,'hourly','mean');

you can also define your own time steps (for example, 15 minute averages):

TimeStep=[OldData.Date(1):minutes(15):OldData.Date(end)];
NewData2=retime(OldData ,TimeStep,'mean');

to make a timetable, use MathWorks help.