How to plot monthly data for 31 years on the same plot?

I have a table with the first column labelled month_Date. It contains the months from Jan 1987 to Jan 2020. The Second column is just a groupcount. The third column is labelled mean_MeanSoilTemperatureAt1MdegCIDWMean and it contains the mean soil temperature for each of the months. I need to make a plot such that the x-axis is labelled with the months and the y-axis labelled soil temperature. Then, I need the lines of the plot to be each years values and for each line to be a specific colour. If anyone could help, it would be much appreciated.

[Leo Tu]

Update: I orginally had a table called T with dates from 01/01/1987 00:00 to 01/01/2020 00:00 and I used the following script to group the dates into months and to find the mean of soil temperature for each month:

TT = groupsummary(T,1,"Month","Mean",38);

where 1 is the column number for the dates in T and 38 is the column number for soil temperature in T.

Then TT is my new table as mentioned in my original question post.

Answer 1

[Seth Furman]

It is probably easier to store the data in a timetable and use retime to get the monthly averages.

https://www.mathworks.com/help/matlab/ref/timetable.html

https://www.mathworks.com/help/matlab/ref/timetable.retime.html

tt = timetable(datetime(2000,1,1)+calmonths(1:100)',(1:100)')

tt = 100×1 timetable
       Time        Var1
    ___________    ____

    01-Feb-2000      1 
    01-Mar-2000      2 
    01-Apr-2000      3 
    01-May-2000      4 
    01-Jun-2000      5 
    01-Jul-2000      6 
    01-Aug-2000      7 
    01-Sep-2000      8 
    01-Oct-2000      9 
    01-Nov-2000     10 
    01-Dec-2000     11 
    01-Jan-2001     12 
    01-Feb-2001     13 
    01-Mar-2001     14 
    01-Apr-2001     15 
    01-May-2001     16 

tt = retime(tt,'monthly','mean')

tt = 100×1 timetable
       Time        Var1
    ___________    ____

    01-Feb-2000      1 
    01-Mar-2000      2 
    01-Apr-2000      3 
    01-May-2000      4 
    01-Jun-2000      5 
    01-Jul-2000      6 
    01-Aug-2000      7 
    01-Sep-2000      8 
    01-Oct-2000      9 
    01-Nov-2000     10 
    01-Dec-2000     11 
    01-Jan-2001     12 
    01-Feb-2001     13 
    01-Mar-2001     14 
    01-Apr-2001     15 
    01-May-2001     16 

Then, we can find the list of years covered by our data ...

years = unique(tt.Time.Year,'sorted')

years = 9×1

2000 2001 2002 2003 2004 2005 2006 2007 2008

... and plot them serially ...

figure
hold on
for i = 1:length(years)
    ttYear = tt(tt.Time.Year == years(i),:);
    plot(ttYear.Time,ttYear.Var1);
end
legend(string(years));
hold off

... or plot them against the months of the year by shifting all the datetime x-values to the same arbitrary year.

figure
hold on
for i = 1:length(years)
    ttYear = tt(tt.Time.Year == years(i),:);
    ttYear.Time.Year = 2000; % set all datetime x-values to the same arbitrary year
    plot(ttYear.Time,ttYear.Var1);
end
xtickformat('MMM');
legend(string(years));
hold off

[Leo Tu]

Thank you @Seth Furman for your help!

[Lejla Latifovic]

Hi @Seth Furman,

Is there a way to plot them against months but not set an arbitrary year so that each line still keeps its corresponding year. Some of the years have the same colours (e.g. above 2000 and 2007).

[Seth Furman]

Do you mean the serial case above (copied below)?

tt = timetable(datetime(2000,1,1)+calmonths(1:100)',(1:100)');
tt = retime(tt,'monthly','mean');
years = unique(tt.Time.Year,'sorted');
f = figure;
hold on
for i = 1:length(years)
    ttYear = tt(tt.Time.Year == years(i),:);
    plot(ttYear.Time,ttYear.Var1);
end
legend(string(years));
hold off

The colors repeat because, by default, only 7 colors are used. After the 7th line, the colors repeat. We can override this behavior by setting our own colors using the colororder function. When setting the color order, we can take advantage of predefined color maps such as parula and adapt them to the number of lines we have.

https://www.mathworks.com/help/matlab/ref/colororder.html

https://www.mathworks.com/help/matlab/colors-1.html#expand_panel_heading_4

https://www.mathworks.com/help/matlab/ref/parula.html

co = parula(numel(years));
colororder(f,co);

[Lejla Latifovic]

Not quite like this, I have nine years of half-hourly data (I've converted it to daily and monthly) I want the x-axis to be months, January - December, and all of the years plotted against the montlhy x-axis. I was looking for a way to do this without setting an arbitrary year for all of the years like you have in the code above but it looks like there isn't another way to do this. Thanks for your response!

Thank you for the colour information, I will attempt this.

Answer 2

[KSSV]

Let T be your table.

thedates = T.(1) ; 
val = T.(3) ; 
[Y,MO,D,H,MI,S] = datevec(thedates) ; 
[c,ia,ib] = unique(Y) ; 
figure(1)
hold on
for i = 1:length(c)
    plot(thedates(ib==i),val(ib==i))
end
legend

[Leo Tu]

Thank you for your response @KSSV. With this script however, I get this plot where the x-axis are labelled in years rather than months and the lines of yearly data are in one line and not overlapping. I have attached the plot for reference. I need the x-axis to be from Jan to Dec and then each yearly data (the coloured lines) to be overlapping. Thank you for your help @KSSV

[Leo Tu]

Sorry @KSSV I made a slight error. when I try using my table (TT) I get an error regarding datevec, stating that the input was was not an array of character vectors or strings. Sorry for the confusion. When I use my orginal table (T) then I get the plot I have shown above.