Extracting and plotting max/min values from timeseries

I have a time series with continuous measurements of pH. The measurements are every 5 minutes, and the timeseries is a total of 13 days.

I want to find and plot the max and min value for each day, where the max and min values are based on 24 observations of a mean pH/hour.

So far I've been able to get the mean pH per hour, but I don't know how to find the max/min values from these and then plot as a continuous time series.

This is what I've done so far to find the mean pH for every hour:

dnv = [time];
dv6 = datevec(dnv);                                     % Date vector matrix, 6 columns
dn4 = datenum([dv6(:,1:4) repmat([0 0],size(dv6,1),1)]);

[Du,~,di] = unique(dn4, 'stable');
Means = accumarray(di, pH, [], @mean);              % Creating means for every hour
Out = [datevec(Du)  Means];                         % Output separated columns, date time mean
de = datetime(Out(:,1:6));                          % Merged date and time back into 1 column

So, how do I find the max and min values with the associated date and time for plotting?

Answer 1

[Peng Li]

Better take advantage of table format and splitapply function.

tbl = readtable('data.csv');

tbl.day  = day(tbl.Date);
tbl.hour = hour(tbl.Date);
 
[Group, hourTbl] = findgroups(tbl(:, {'day', 'hour'}));
hourTbl.mean = splitapply(@nanmean, tbl.pH, Group);

[Group, dayTbl]  = findgroups(hourTbl(:, 'day'));

dayTbl.max = splitapply(@max, hourTbl.mean, Group);
dayTbl.min = splitapply(@min, hourTbl.mean, Group);

[Lindazeta]

That's awesome for getting the max and min values, but I still need the time stamp that goes with the values so I can plot them as a time series. Or am I missing something in your code?

[Peng Li]

Modified a bit to show max pH and the corresponding max time. It's a bit twisted but it's working.

tbl = readtable('data.csv');
tbl.year  = year(tbl.Date);
tbl.month = month(tbl.Date);
tbl.day   = day(tbl.Date);
tbl.hour  = hour(tbl.Date);
 
[Group, hourTbl] = findgroups(tbl(:, {'year', 'month', 'day', 'hour'}));
hourTbl.mean = splitapply(@nanmean, tbl.pH, Group);

[Group, dayTbl]  = findgroups(hourTbl(:, {'year', 'month', 'day'}));
temp = splitapply(@(x, y) mymax(x, y), hourTbl.mean, hourTbl.hour, Group);
dayTbl.maxHour   = temp(:, 2);
dayTbl.maxpH     = temp(:, 1);

dayTbl.time      = datetime(dayTbl.year, dayTbl.month, dayTbl.day, dayTbl.maxHour, 0, 0);

function out = mymax(ph, time)
[m, ind] = max(ph);
out = [m, time(ind)];
end

[Lindazeta]

Awesome! Thanks! One last question though... How do I add the min values? I tried re-writing it but my matlab skills are not great...

[Lindazeta]

Oh never mind, I made it work! Thank you so much for your help!