How do you plot data bi-monthly for the course of a year?

I have a data for a year where I need two plots for each month covering the 1st-15th and 16th-end of month, i.e. bi-monthly plots of this data.

The dataset isn't entirely complete and at some points there is no data for the 1st/15th/16th (as well as other days) of a month.

The data is currently in a timetable with time/date in datetime format 'dd/MM/yyyy HH:mm:ss'.

Is there a way of looping through the data to produce these bi-monthly plots?

My first instinct was to use Timerange, but cannot work out if there is a way to define 'half a month' using this function and plotting every 2 weeks or 15 days soon gets out of phase.

Below is a basic example of what the data looks like (I have added ... just to imply data continues during that time period)

     date              data
01/01/2018 00:00:00   344.98
01/01/2018 00:30:00   304.13
01/01/2018 01:00:00   355.74
...
13/01/2018 09:30:00   554.13
18/01/2018 10:00:00   344.77
19/01/2018 10:00:00   346.47
27/01/2018 00:30:00   304.13
02/02/2018 01:00:00   357.83
02/02/2018 01:30:00   367.87
...
31/12/2018 22:30:00	634.63
31/12/2018 23:00:00	400.53
31/12/2018 23:30:00	621.43

[Dom Smith]

Perhaps the easiest solution will be to adjust the plot limits

with xlim(datetime(2018,[1 1],[1 15]))

and xlim(datetime(2018,[1 16],[1 30])) and so on (adjusting 30 depending on the number of days in a month).

But I am still not sure how I would write this into a loop.

Answer 1

[A. Sawas]

You may use the function datetime to set the lower and upper dates of each bi-monthly period. Then using find to selectively choose the data points between them.

y = 2018;
for m=1:12
	t1 = datetime(y,m,01);
	t2 = datetime(y,m,16);
	i = find(DataT.date => t1 && DataT.date < t2);
      % Plot the first half of the month
	plot(DataT.date(i), DataT.data(i));

	t1 = t2;
	t2 = datetime(y,m+1,1); % when m=12 the returned date is the first day of the next year
	i = find(DataT.date => t1 && DataT.date < t2);
      % Plot the second half of the month
	plot(DataT.date(i), DataT.data(i));
end

[Dom Smith]

Thanks very much, with a bit of tweaking and correction that was the basis for a nice bit of code.

[Carol pacheco]

and for cases of wanting to separate a large dataset? where the number of weeks are not exact (23 weeks and 3 days for example ?

[A. Sawas]

You can handle the fractions in a separate block of code