Plotting Day and Time from csv file

I am attempting to plot hourly power data from the first day of each month for all 24 hours of that day for one year (288 date/time points). Ideally, it should be grouped into twelve curves, one curve for each day. I extracted the date-time information from a .csv file and plotted my output power data against these date-time values. Unfortunately, MATLAB automatically filled in all the days in between the first day of each month with zeros on my graph, so I have twelve sharp, unreadable peaks instead of twelve curves where I can read the data at each hour.

The code I used to plot is attached, as well as the hourlyTempWindData.csv, Q_Total is a 1x288 double that's calculated by the rest of my code (which is very long), so I've included a random number array instead. I want my graph to look like the second one (one day with a data point at each hour) instead of the first (unreadable peak with flat space in between).

Also, is there anything I can do to stop the orange warning from showing up after I use readtable to get the data from the csv? It isn't interfering with my code running, it's just annoying.

%Import radiation and weather data sheets, and set variables for each
%column, then transpose values to a row vector

hourlyData = readtable("hourlyTempWindData.csv", 'VariableNamingRule','preserve')

hourlyData = 288×4 table
           DATE            HourlyDryBulbTemperature    HourlyWindSpeed    HourlyWindSpeed ft/s
    ___________________    ________________________    _______________    ____________________

    2019-01-01T00:54:00               10                     21                    30.8       
    2019-01-01T01:54:00                9                     22                  32.267       
    2019-01-01T02:54:00                7                     20                  29.333       
    2019-01-01T03:54:00                7                     15                      22       
    2019-01-01T04:54:00                6                     18                    26.4       
    2019-01-01T05:54:00                5                     18                    26.4       
    2019-01-01T06:54:00                5                     21                    30.8       
    2019-01-01T07:54:00                5                     15                      22       
    2019-01-01T08:54:00                5                     14                  20.533       
    2019-01-01T09:54:00                6                     13                  19.067       
    2019-01-01T10:54:00                6                     11                  16.133       
    2019-01-01T11:54:00                7                     10                  14.667       
    2019-01-01T12:54:00               10                     10                  14.667       
    2019-01-01T13:54:00               12                      5                  7.3333       
    2019-01-01T14:54:00               14                      6                     8.8       
    2019-01-01T15:54:00               15                      3                     4.4       

dates = datetime(table2array(hourlyData(:,1)));

Q_Total = randi(50, [1 288]);

%Graph daily energy losses/gains against date and hour, results in 12 sharp
%peaks with flat space between

figure(1)
hold off
grid on
plot(dates, Q_Total)
title('Hourly Energy Needs for Greenhouse');
xlabel('Date and Time');
ylabel('Q (Btu/h)');

%Graph daily energy losses/gains, results in 12 curves (when using actual
%data, not randi)

figure(2)
hold off
grid on
plot(Q_Total);
title('Hourly Energy Needs for Greenhouse');
xlabel('Date and Time');
ylabel('Q (Btu/h)');

Answer 1

[Cris LaPierre]

With your data as datetimes, there is no good way to crop out the missing days, and workarounds I might try are cumbersome. Since the ideal is to have all 12 days plotted as their own line, I'll do that.

hourlyData = readtable("hourlyTempWindData.csv", 'VariableNamingRule','preserve');

% add power, month (for grouping) and 
hourlyData.Q_Total = randi(50, [288,1]);
hourlyData.Month = month(hourlyData.DATE);
hourlyData.Hour = hour(hourlyData.DATE);

G = findgroups(hourlyData.Month);

figure
hold on
splitapply(@(x,y) plot(x,y),hourlyData(:,["Hour","Q_Total"]),G)
hold off
ylabel('Power')
xlabel('Hour of Day')
legend(num2str(unique(hourlyData.Month)))

[Aussia S.]

This works, but I'm hoping to plot the January values next to the February ones, etc. sequentially so the graph looks like the attached, but the date and time labels are on the x-axis.

[Cris LaPierre]

" Ideally, it should be grouped into twelve curves, one curve for each day"

I misunderstood that to mean one line for each day. Ok, so one line with a spike for each day? How did you create the attached plot?

[Aussia S.]

plot(Q_Total)

[Cris LaPierre]

So what is the vision? You can't label all 288 points. That will be illegible. But if you only label some (say every 50 points like in the plot you've attached), something is lost. I don't think there is a great way of doing this, but here's what I could think of on the spot.

Basically, turn the DATE variable into strings, and then relabel the xticklabels with the corresponding string.

hourlyData = readtable("hourlyTempWindData.csv", 'VariableNamingRule','preserve');

% Convert DATE to strings so they can be used to label X-Axis
% I selected Month Hour for the format, but it can be modified
Dates = string(hourlyData.DATE,"MMM HH");

% compute power (done by the rest of your code)
Q_Total = randi(50, [288,1]);

plot(Q_Total);
title('Hourly Energy Needs for Greenhouse');
xlabel('Date and Time');
ylabel('Q (Btu/h)');

% Set the xlims to the range of the data so that the tick locations are
% valid indices into the Dates variable
xlim([1 height(hourlyData)])
% Extract current tick locatoins
xt = xticks;
% Relabel xticks with corresponding string from Dates
xticklabels(Dates(xt))

[Aussia S.]

My ultimate vision was to lable each point with its respective time, but I understand this would get messy very quickly, as you noted. Can the xticks be changed so each peak is labeled with its respective month? Perhaps centering the xtick where the peak of each day is (or at least close to that general area. Thank you for your help!

[Cris LaPierre]

Yes. Use the same format as what I've shared. The only additions are

• Find each peak (options include max, findpeaks, or Find Local Extrema live task)
• Extract the index of each peak (returned by the functions I mentioned)
• Use xticks to set the tick locations to the desired values

There rest is already there for you.