How can I divide a week of a timetable into hours

Hi,

is it possible to devide a week of a timetable into 168 hours?

I would need the means of a timetable from monday to sunday on an hourly basis so I can show how its behaving of the week

Thx :)

[Andrei Bobrov]

Please attach your timetable as mat-file.

[Antonio Melieni]

So what I need is the mean over each of the 168 hours of the week from these data

Answer 1

[Guillaume]

hourlytimetable = retime(yourtimetable, 'hourly', 'mean')

[Antonio Melieni]

Thx, thats what I knew so far. But I need the whole week devided in 168 hours as I want to show the behaviour of variable on an hourly basis over the week. Is that possible?

[Guillaume]

I'm not sure I understand, the above divides all the weeks on an hourly basis.

Oh, maybe you just want 168 bins regardless of the number of weeks. I.e. bin 1 is the first hour of each week.

groupsummary has many binning options but 'hourofweek' is not an option, so you'll have to calculate the bins yourself. Not too difficult:

yourtimetable.bin = (day(yourtimetable.Time, 'dayofweek')-1)*24 + hour(yourtimetable.Time);
groupsummary(yourtimetable, 'bin', 'mean')

[Antonio Melieni]

Ok. I have one year tick data of returns. I need a plot where the behaviour of the returns is shown on an hourly basis of the week. Do you know what I mean?

[Antonio Melieni]

I attached an example of what I need. Please have a look at the PDF.

[Guillaume]

It seems that my groupsummary answer above gives you your intra-week data (once you've fixed the 7 that should have been 24, -now edited-),as for the intra-day data, it's even simpler

intraday = groupsummary(Volatility, 'Time', 'hourofday', 'mean')

and so, using correct variable names:

Volatility.HourOfWeek = (day(Volatility.Time, 'dayofweek')-1)*24 + hour(Volatility.Time);
intraweek = groupsummary(Volatility, 'HourOfWeek', 'mean')

[Antonio Melieni]

THANK YOU SOOOOOOOOOOOOOOOOOOO MUCH!!!!!!!

[Peter Perkins]

Guillaume is correct.

retime can do hourly (or whatever) summaries, but only on a linear time line. You are asking for a cyclic time line. Adding a dummy variable for "hour of week" and grouping on that is the way to go. groupsummary, or splitapply, or varfun would allwork; groupsummary is probably the simplest.

[Guillaume]

Peter, a suggestion: an option to specify bin indices instead of bin intervals for groupsummary at least. That would avoid having to create the binning variable:

hourofweek = ...
groupsummary(Volatility, 'BinNumbers', hourofweek, 'mean')

[Peter Perkins]

Yes, this would be useful in many contexts. Thanks.