I have made some modifications. The heatmap below Shows highest temp. each hour of a month, for two years.
% creating table (i hope you are using new Version of matlab
% which has table function, otherwise try dataset)
t1 = datetime(2016,01,1,8,0,0);
t2 = datetime(2018,01,1,8,0,0);
t = t1:hours(1):t2;
myTable = table(t',year(t'),month(t'),day(t'),hour(t'));
myTable.Properties.VariableNames = {'date' 'year' 'month' 'day' 'hour'};
a = 0;
b = 40;
r = (b-a).*rand(17545,1) + a;
myTable.temp = r;
% Create categorical arrays from the month
% and hour columns of the table.
% Then determine the unique months and hours
% to use as labels along the x-axis and y-axis.
months = categorical(myTable.month);
hours = categorical(myTable.hour);
xlabels = categories(months);
ylabels = categories(hours);
%Determine the final size of the resulting
%color data based on the number of unique months and hours.
nummonths = numel(xlabels);
numhours = numel(ylabels);
%Convert the categorical months and hours arrays
%into numeric indices to use with the accumarray function.
%Compute the color data as the maximum temperature for each
%month and hour combination using the accumarray function.
%Use NaN for missing month and year combinations.
x = double(months);
y = double(hours);
temps = myTable.temp;
cdata = accumarray([y,x],temps,[numhours,nummonths],@max,NaN);
%Create the heatmap. Label the x-axis and y-axis with the
%months and years, respectively. Color the heatmap cells
%using the computed matrix data.
h = heatmap(xlabels,ylabels,cdata);
