histogram of datetimes with customisable bin width

How to plot a histogram of (differences of) datetimes, with a customizable bin width, such as 1 minute ?

This is my attempt, but something is wrong/missing.....

% input
a = datetime({'2022-05-17 16:07:28.280'
   '2022-05-17 16:08:55.968'
   '2022-05-17 16:19:26.632'
   '2022-05-17 16:21:23.840'
   '2022-05-17 16:22:47.540'
   '2022-05-17 16:25:01.152'
   '2022-05-17 16:26:27.728'
   '2022-05-17 16:32:47.192'
   '2022-05-17 18:00:01.340'
   '2022-05-17 18:05:41.080'
   '2022-05-17 18:07:11.248'
   '2022-05-17 18:08:59.208'
   '2022-05-17 18:12:17.940'
   '2022-05-17 18:19:33.520'
   '2022-05-17 18:21:10.020'
   '2022-05-17 18:22:20.740'
   '2022-05-17 18:23:55.900'
   '2022-05-17 18:25:05.460'
   '2022-05-17 18:29:46.080'
   '2022-05-17 20:02:22.160'
   '2022-05-17 20:03:38.900'
   '2022-05-17 20:05:17.112'
   '2022-05-17 20:19:10.072'
   '2022-05-17 20:20:38.928'
   '2022-05-17 20:21:41.000'
   '2022-05-17 20:23:16.260'
   '2022-05-17 20:24:27.460'
   '2022-05-17 20:28:48.980'});

% histcounts 
b = diff(a);
binLimits = [min(b) max(b)];
binWdith = datetime('00:01:00');
numBins = ceil(minutes(diff(binLimits))/minute(binWdith));
[counts,binEdges]=histcounts(b, 'BinLimits',binLimits,'BinWidth',binWdith);

Error using datetime/histcounts
X must be a datetime array.

% histogram
histogram(b,'binEdges',binEdges,'FaceColor','k','EdgeColor','k');

Answer 1

[Star Strider]

After doing a few conversions between datetime and duration to get all the arguments to be what the functions want —

% input
a = datetime({'2022-05-17 16:07:28.280'
   '2022-05-17 16:08:55.968'
   '2022-05-17 16:19:26.632'
   '2022-05-17 16:21:23.840'
   '2022-05-17 16:22:47.540'
   '2022-05-17 16:25:01.152'
   '2022-05-17 16:26:27.728'
   '2022-05-17 16:32:47.192'
   '2022-05-17 18:00:01.340'
   '2022-05-17 18:05:41.080'
   '2022-05-17 18:07:11.248'
   '2022-05-17 18:08:59.208'
   '2022-05-17 18:12:17.940'
   '2022-05-17 18:19:33.520'
   '2022-05-17 18:21:10.020'
   '2022-05-17 18:22:20.740'
   '2022-05-17 18:23:55.900'
   '2022-05-17 18:25:05.460'
   '2022-05-17 18:29:46.080'
   '2022-05-17 20:02:22.160'
   '2022-05-17 20:03:38.900'
   '2022-05-17 20:05:17.112'
   '2022-05-17 20:19:10.072'
   '2022-05-17 20:20:38.928'
   '2022-05-17 20:21:41.000'
   '2022-05-17 20:23:16.260'
   '2022-05-17 20:24:27.460'
   '2022-05-17 20:28:48.980'});

% histcounts 
b = diff(a);
dtb = datetime(string(b));
binLimits = [min(dtb) max(dtb)];
binWdith = duration('00:01:00');
numBins = ceil(minutes(diff(binLimits))/minutes(binWdith));
[counts,binEdges]=histcounts(dtb, 'BinLimits',binLimits,'BinWidth',binWdith);
% histogram
histogram(dtb,'binEdges',binEdges,'FaceColor','k','EdgeColor','k');

Is that the desired result?

.

[Sim]

thanks a lot!!!!

I would like to accept both answers @Star Strider @Cris LaPierre !!

[Cris LaPierre]

Since b is created already using the datetime function, the following step is not necessary

dtb = datetime(string(b));

[Sim]

thanks a lot @Cris LaPierre ! :-)

Answer 2

[Cris LaPierre]

By taking the diff of a, your variable b is now a duration, not a datetime. That means your binWidth must be of the same datatype. Make that a duration, and you code works.

I didn't see a need for numBins, so I removed that line of code.

% input
a = datetime({'2022-05-17 16:07:28.280'
   '2022-05-17 16:08:55.968'
   '2022-05-17 16:19:26.632'
   '2022-05-17 16:21:23.840'
   '2022-05-17 16:22:47.540'
   '2022-05-17 16:25:01.152'
   '2022-05-17 16:26:27.728'
   '2022-05-17 16:32:47.192'
   '2022-05-17 18:00:01.340'
   '2022-05-17 18:05:41.080'
   '2022-05-17 18:07:11.248'
   '2022-05-17 18:08:59.208'
   '2022-05-17 18:12:17.940'
   '2022-05-17 18:19:33.520'
   '2022-05-17 18:21:10.020'
   '2022-05-17 18:22:20.740'
   '2022-05-17 18:23:55.900'
   '2022-05-17 18:25:05.460'
   '2022-05-17 18:29:46.080'
   '2022-05-17 20:02:22.160'
   '2022-05-17 20:03:38.900'
   '2022-05-17 20:05:17.112'
   '2022-05-17 20:19:10.072'
   '2022-05-17 20:20:38.928'
   '2022-05-17 20:21:41.000'
   '2022-05-17 20:23:16.260'
   '2022-05-17 20:24:27.460'
   '2022-05-17 20:28:48.980'});

% histcounts 
b = diff(a);
binLimits = [min(b) max(b)];
binWidth = duration('00:01:00');

[counts,binEdges]=histcounts(b, 'BinLimits',binLimits,'BinWidth',binWidth);
% histogram
histogram(b,'binEdges',binEdges,'FaceColor','k','EdgeColor','k');

[Sim]

thanks a lot!!!!

I would like to accept both answers @Star Strider @Cris LaPierre !!

(yes, "numBins" is not useful here.. :-) )

[Cris LaPierre]

numBins is also just the length of counts.

[Sim]

thanks :)