How can I plot this date time graph?

Hi!

I have this Years.mat timetable where I have three columns - a) date, b) hours of the date, and c) tide height.

Can anyone please help me to know how can I plot a graph where I can show all the months in x-axis and the corresponding heights of those months in the y-axis? Just like this picture -

Any feedback will be much helpful!!

[Walter Roberson]

I think it would make more sense as a scatter plot than as a line plot.

[Ashfaq Ahmed]

Yes, I agree. Can you please give me an idea on the scatter plot of this problem?

Answer 1

[Sulaymon Eshkabilov]

Here is how you can get the plot:

load('Years.mat')
N = numel(Climatology.HeightSeries);
plot(1:N,Climatology.HeightSeries), shg
xticks(linspace(1, N, 12));
xticklabels({'Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun', 'Jul', 'Aug', 'Sep', 'Oct', 'Nov', 'Dec'})
xlabel('Date Time')
ylabel('Tide Height, (m)')
title('Monthly variation of tide height (2004-2021)')

[Edit to run the code and show the plot. -Voss]

[Ashfaq Ahmed]

excellent!

[Voss]

@Ashfaq Ahmed: Note that the plot generated by this answer is the same as the plot generated by the other answer; this answer merely labels the x-axis with the names of the months, which is misleading.

For example:

• the x-tick labeled 'Jan' here actually corresponds to data from the beginning of 2004
• the x-tick labeled 'May' here actually corresponds to data from around 2009-2010
• the x-tick labeled 'Sep' here actually corresponds to data from around 2015-2016
• the x-tick labeled 'Dec' here actually corresponds to data from the beginning of 2022
• etc.

I don't think this is the result you're after.

[Ashfaq Ahmed]

Oh yes. You are right!! Thank you so much for the check. I was about to put the plot in my paper.

[Sulaymon Eshkabilov]

Here some addional coding is required.

Step (1). Sort data by all the dates (only data collected ones) and combine the data by dates

Step (2) Average the values per date for every day (data collected ones) of each month of the year

Step (3) Then plot the data by dates (month) per year

Answer 2

[Les Beckham]

load('Years.mat')
whos

  Name                  Size               Bytes  Class        Attributes

  Climatology      180468x2             23462081  timetable              
  ans                   1x34                  68  char                   
  cmdout                1x33                  66  char                   

Climatology

Climatology = 180468×2 timetable
    TimeSeries     HourSeries    HeightSeries
    ___________    __________    ____________

    01-Jan-2004    {'00:00'}         0.207   
    01-Jan-2004    {'01:00'}         0.189   
    01-Jan-2004    {'01:12'}         0.226   
    01-Jan-2004    {'02:00'}         0.263   
    01-Jan-2004    {'03:00'}         0.318   
    01-Jan-2004    {'04:00'}         0.394   
    01-Jan-2004    {'05:00'}          0.58   
    01-Jan-2004    {'06:00'}         0.843   
    01-Jan-2004    {'07:00'}         1.065   
    01-Jan-2004    {'08:00'}         1.149   
    01-Jan-2004    {'08:12'}        1.1155   
    01-Jan-2004    {'09:00'}         1.082   
    01-Jan-2004    {'10:00'}         0.918   
    01-Jan-2004    {'11:00'}         0.693   
    01-Jan-2004    {'12:00'}         0.458   
    01-Jan-2004    {'13:00'}         0.322   

t = Climatology.TimeSeries + hour(Climatology.HourSeries);
scatter(t, Climatology.HeightSeries)
grid on
xlabel('Time')
ylabel('Tide Height (m)')

I actually think plot looks better than scatter

figure
plot(t, Climatology.HeightSeries)
grid on
xlabel('Time')
ylabel('Tide Height (m)')

[Ashfaq Ahmed]

Hi! Thank you so much! But I actually want to keep the months in the x-axis. Not the year!

[Walter Roberson]

load('Years.mat')
t = Climatology.TimeSeries + hour(Climatology.HourSeries);
[y, m, d] = ymd(t);
y(:) = min(y);
hacked_t = datetime(y, m, d);
scatter(hacked_t, Climatology.HeightSeries)
grid on
xlabel('Time')
ylabel('Tide Height (m)')

That is, we hack all of the dates replacing the year with the earliest year, so that (say) May 19th of one year is the same relative offset as May 19th of a different year.

Other approaches have a challenge in trying to get the month label right... unless you want to convert to day of year and use datetick()... and that has challenges with leap years.

[Voss]

load Years
Climatology

Climatology = 180468×2 timetable
    TimeSeries     HourSeries    HeightSeries
    ___________    __________    ____________

    01-Jan-2004    {'00:00'}         0.207   
    01-Jan-2004    {'01:00'}         0.189   
    01-Jan-2004    {'01:12'}         0.226   
    01-Jan-2004    {'02:00'}         0.263   
    01-Jan-2004    {'03:00'}         0.318   
    01-Jan-2004    {'04:00'}         0.394   
    01-Jan-2004    {'05:00'}          0.58   
    01-Jan-2004    {'06:00'}         0.843   
    01-Jan-2004    {'07:00'}         1.065   
    01-Jan-2004    {'08:00'}         1.149   
    01-Jan-2004    {'08:12'}        1.1155   
    01-Jan-2004    {'09:00'}         1.082   
    01-Jan-2004    {'10:00'}         0.918   
    01-Jan-2004    {'11:00'}         0.693   
    01-Jan-2004    {'12:00'}         0.458   
    01-Jan-2004    {'13:00'}         0.322   

Seems like this:

t = Climatology.TimeSeries + hour(Climatology.HourSeries)

t = 180468×1 datetime array
   01-Jan-2004
   02-Jan-2004
   02-Jan-2004
   03-Jan-2004
   04-Jan-2004
   05-Jan-2004
   06-Jan-2004
   07-Jan-2004
   08-Jan-2004
   09-Jan-2004
   09-Jan-2004
   10-Jan-2004
   11-Jan-2004
   12-Jan-2004
   13-Jan-2004
   14-Jan-2004
   14-Jan-2004
   15-Jan-2004
   16-Jan-2004
   17-Jan-2004
   18-Jan-2004
   19-Jan-2004
   20-Jan-2004
   21-Jan-2004
   21-Jan-2004
   22-Jan-2004
   23-Jan-2004
   24-Jan-2004
   02-Jan-2004
   03-Jan-2004

Should be this:

t = Climatology.TimeSeries + duration(Climatology.HourSeries,'InputFormat','hh:mm')

t = 180468×1 datetime array
   01-Jan-2004 00:00:00
   01-Jan-2004 01:00:00
   01-Jan-2004 01:12:00
   01-Jan-2004 02:00:00
   01-Jan-2004 03:00:00
   01-Jan-2004 04:00:00
   01-Jan-2004 05:00:00
   01-Jan-2004 06:00:00
   01-Jan-2004 07:00:00
   01-Jan-2004 08:00:00
   01-Jan-2004 08:12:00
   01-Jan-2004 09:00:00
   01-Jan-2004 10:00:00
   01-Jan-2004 11:00:00
   01-Jan-2004 12:00:00
   01-Jan-2004 13:00:00
   01-Jan-2004 13:48:00
   01-Jan-2004 14:00:00
   01-Jan-2004 15:00:00
   01-Jan-2004 16:00:00
   01-Jan-2004 17:00:00
   01-Jan-2004 18:00:00
   01-Jan-2004 19:00:00
   01-Jan-2004 20:00:00
   01-Jan-2004 20:48:00
   01-Jan-2004 21:00:00
   01-Jan-2004 22:00:00
   01-Jan-2004 23:00:00
   02-Jan-2004 00:00:00
   02-Jan-2004 01:00:00

[Walter Roberson]

@Voss is correct that @Sulaymon Eshkabilov and @Les Beckham do not combine data across years.

My hack-the-year approach does combine across years.

[Sulaymon Eshkabilov]

In fact, this is a quite tricky exercise and there are many repetivie dates. So far, no answer gives good sorted or combined time series (@Walter Roberson, @Voss). E.g. Walter's answer:

load('Years.mat')
t = Climatology.TimeSeries + hour(Climatology.HourSeries);
[y, m, d] = ymd(t);
y(:) = min(y);
hacked_t = datetime(y, m, d);
hacked_t(611)  

ans = datetime
   31-Jan-2004

hacked_t(612)

ans = datetime
   01-Feb-2004

hacked_t(629)

ans = datetime
   24-Jan-2004

hacked_t(639)

ans = datetime
   01-Feb-2004

hacked_t(657)

ans = datetime
   25-Jan-2004

hacked_t(711)

ans = datetime
   27-Jan-2004

hacked_t(743)

ans = datetime
   01-Feb-2004

hacked_t(763)

ans = datetime
   18-Feb-2004

hacked_t(793)

ans = datetime
   30-Jan-2004

% In some dates, 26 samples, and in others 27 or 28 or .. data points were collected
Climatology.HeightSeries(Climatology.TimeSeries=='31-Jan-2004') 

ans = 27×1

0.2390 0.1930 0.2850 0.3940 0.4790 0.6110 0.8020 0.9660 0.9975 1.0290

numel(Climatology.HeightSeries(Climatology.TimeSeries=='30-Jan-2004'))

ans = 27

numel(Climatology.HeightSeries(Climatology.TimeSeries=='01-Feb-2004'))

ans = 26

% Therefore, none of the combined and sorted time series is accurate.
% Here is:

plot(hacked_t(Climatology.TimeSeries=='30-Jan-2004'), Climatology.HeightSeries(Climatology.TimeSeries=='30-Jan-2004'))

[Walter Roberson]

I assumed that Les's calculation of time was correct, but it had an error. hour() of a duration returns an integer, and when added to a datetime that integer is treated as days unless you specifically convert it to duration

load('Years.mat')
t = Climatology.TimeSeries + hours(hour(Climatology.HourSeries));
[y, m, d] = ymd(t);
y(:) = min(y);
hacked_t = datetime(y, m, d);
scatter(hacked_t, Climatology.HeightSeries)
grid on
xlabel('Time')
ylabel('Tide Height (m)')

[Les Beckham]

Good catch, Walter.

[Sulaymon Eshkabilov]

Here is a bit more accurately sorted data and averaged data by month.

load('Years.mat')
MONTH_ALL=datetime(Climatology.TimeSeries, 'Format','MMM-uuuu');
CLIMAT = table(MONTH_ALL);
CLIMAT.HS = Climatology.HeightSeries;
plot(CLIMAT.MONTH_ALL, CLIMAT.HS)

t=datetime(2004,01,01):calweeks(4):datetime(2004,12,31);
t = t(2:13); 
m = month(t, 'shortname');
%%  Separates data by month per year for 2004 to 2021, and computes monthly averages  
for ii = 1:12
    for jj=1:(2021-2004)
    Months{ii, jj} = [m{ii} '-' num2str(2003+jj)];
    ANNUAL{ii, jj}=CLIMAT.HS(CLIMAT.MONTH_ALL==Months{ii,jj});
    ANNUAL_ave{ii, jj} = mean(ANNUAL{ii,jj});
    end
end
%% This part is for plotting separated data by month and by year.
% This part is a bit boring code and not elagant or efficient code due to
% data collection inconsistency by month.
% Plot's xticklabels is not crisp with the data. Some some minor adjustment is left out.   
figure('name', 'ALL')
MColor = {'r', 'g', 'b', 'm', 'b', 'c', 'y', 'k', 'r', 'g', 'b', 'm','b', 'c', 'y', 'k','g'};
N=2021-2004;
% January
for jj=1:N
    plot(1:numel(ANNUAL{1, jj}),ANNUAL{1, jj}, '*--','Color',MColor{1}), hold on
end
% February
M2 = numel(ANNUAL{1, N});
for jj=1:N   
    plot(1+M2:numel(ANNUAL{2, jj})+M2,ANNUAL{2, jj},'o-.','Color', MColor{2})    
end
% March
M3 = numel(ANNUAL{1, N})+M2;
for jj=1:N   
    plot(1+M3:numel(ANNUAL{3, jj})+M3,ANNUAL{3, jj},'s-.','Color', MColor{3})    
end
% April
M4 = numel(ANNUAL{1, N})+M3;
for jj=1:N   
    plot(1+M4:numel(ANNUAL{4, jj})+M4,ANNUAL{4, jj},'<-.','Color', MColor{4})    
end
% May
M5 = numel(ANNUAL{1, N})+M4;
for jj=1:N   
    plot(1+M5:numel(ANNUAL{5, jj})+M5,ANNUAL{5, jj},'>-.','Color', MColor{5})    
end
% June
M6 = numel(ANNUAL{1, N})+M5;
for jj=1:N   
    plot(1+M6:numel(ANNUAL{6, jj})+M6,ANNUAL{6, jj},'p-.','Color', MColor{6})    
end
% July
M7 = numel(ANNUAL{1, N})+M6;
for jj=1:N   
    plot(1+M7:numel(ANNUAL{7, jj})+M7,ANNUAL{7, jj},'p-.','Color', MColor{7})    
end
% August
M8 = numel(ANNUAL{1, N})+M7;
for jj=1:N   
    plot(1+M8:numel(ANNUAL{8, jj})+M8,ANNUAL{8, jj},'p-.','Color', MColor{8})    
end
% September
M9 = numel(ANNUAL{1, N})+M8;
for jj=1:N   
    plot(1+M9:numel(ANNUAL{9, jj})+M9,ANNUAL{9, jj},'p-.','Color', MColor{9})    
end
% October
M10 = numel(ANNUAL{1, N})+M9;
for jj=1:N   
    plot(1+M10:numel(ANNUAL{10, jj})+M10,ANNUAL{10, jj},'p-.','Color', MColor{10})    
end
% November
M11 = numel(ANNUAL{1, N})+M10;
for jj=1:N   
    plot(1+M11:numel(ANNUAL{11, jj})+M11,ANNUAL{11, jj},'p-.','Color', MColor{11})    
end
% December
M12 = numel(ANNUAL{1, N})+M11;
for jj=1:N   
    plot(1+M12:numel(ANNUAL{12, jj})+M12,ANNUAL{12, jj},'p-.','Color', MColor{12})    
end
xticks(linspace(1, numel(ANNUAL{12, N})+M12, 12))
xticklabels({'Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun', 'Jul', 'Aug', 'Sep', 'Oct', 'Nov', 'Dec'})
xtickangle(60)
xlim([1, numel(ANNUAL{12, N})+M12])
xlabel('Date Time by month for 2004 - 2021')
ylabel('Tide Height, (m)')
title('Monthly variation of tide height (2004-2021)')
hold off

%%  Averaged per month for 2004 - 2021
figure('name', 'ALL')
MType = {'*', 'o', 's', '<', '>', 'd', 'p', 'h', 's', 'h', 'x', '^'};
MColor ={'r', 'g', 'b', 'm', 'b', 'g', 'b', 'k', 'r', 'k', 'b', 'm'};
N=2021-2004;
for k=1:12
        AVE=cell2mat(ANNUAL_ave(k,1:17));
        plot(k*ones(1,N),AVE, MType{k},'Color',MColor{k}, 'MarkerFaceColor', 'y'), hold on
end
xlim([0,13])
xticks(1:12)
xticklabels({'Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun', 'Jul', 'Aug', 'Sep', 'Oct', 'Nov', 'Dec'})
xtickangle(45)
xlabel('Date Time by month for 2004 - 2021')
ylabel('Tide Height, (m)')
title('Monthly averaged variation of tide height (2004-2021)')
grid on
hold off