Day of Week data plot

Hi, I need to plot daily element concentration averages by day of week (using several months of data). The code I've included does work (I've also included an image of the plot output), but since we measure so many species I'd like to include some kind of for loop to make the process quicker and be able to plot more elements at a time without so much individual specification. Our instrument output gives us concentration and uncertainty columns for each element. Would love any input on this - relatively new to Matlab/coding. Thanks!

Dayofweek = weekday(Xactdata.Date(Data));
Sun = find(Dayofweek == 1); Mon = find(Dayofweek == 2); Tue = find(Dayofweek == 3); Wed = find(Dayofweek == 4);  Thu = find(Dayofweek == 5);  Fri = find(Dayofweek == 6);  Sat = find(Dayofweek == 7);
wk = [1 2 3 4 5 6 7];

S_wk = [Xactdata.SULPHUR(Data)]; 
S_wk_err = [Xactdata.S_U(Data)];
Cl_wk = [Xactdata.CHLORINE(Data)]; 
Cl_wk_err = [Xactdata.Cl_U(Data)];

Xact_week_Cl = [mean(Cl_wk(Sun)) mean(Cl_wk(Mon)) mean(Cl_wk(Tue)) mean(Cl_wk(Wed)) mean(Cl_wk(Thu)) mean(Cl_wk(Fri)) mean(Cl_wk(Sat))];
Xact_err_week_Cl = [mean(Cl_wk_err(Sun)) mean(Cl_wk_err(Mon)) mean(Cl_wk_err(Tue)) mean(Cl_wk_err(Wed)) mean(Cl_wk_err(Thu)) mean(Cl_wk_err(Fri)) mean(Cl_wk_err(Sat))];
Xact_week_S = [mean(S_wk(Sun)) mean(S_wk(Mon)) mean(S_wk(Tue)) mean(S_wk(Wed)) mean(S_wk(Thu)) mean(S_wk(Fri)) mean(S_wk(Sat))];
Xact_err_week_S = [mean(S_wk_err(Sun)) mean(S_wk_err(Mon)) mean(S_wk_err(Tue)) mean(S_wk_err(Wed)) mean(S_wk_err(Thu)) mean(S_wk_err(Fri)) mean(S_wk_err(Sat))];

figure
errorbar(wk, Xact_week_S, Xact_err_week_S)
hold on 
errorbar(wk, Xact_week_Cl, Xact_err_week_Cl)
xticklabels({'Sun', 'Mon', 'Tue', 'Wed', 'Thu', 'Fri', 'Sat'})
title('Weekly Concentrations'); %Modify as needed
legend('S', 'Cl')
ylabel('Concentration (ng/m^3)');

Answer 1

[Star Strider]

Perhaps something like this —

Date = datetime(2023,1,1, 'Format','yyyy-MM-dd')+caldays(0:365).';
Cl = randn(size(Date)) + 30*sin(2*pi*(0:numel(Date)-1)/6).' + 60;
S = randn(size(Date)) + 60;
F = randn(size(Date)) + 30*cos(2*pi*(0:numel(Date)-1)/6).' + 60;

Data = table(Date,Cl,S,F)

Data = 366×4 table
       Date         Cl        S         F   
    __________    ______    ______    ______

    2023-01-01    60.404    59.843    90.561
    2023-01-02    84.582    59.642    74.518
    2023-01-03    86.675    57.159    44.085
    2023-01-04    61.076      59.3    31.712
    2023-01-05      32.1     61.74    44.652
    2023-01-06     35.14    59.217    76.156
    2023-01-07    57.455    59.461    92.111
    2023-01-08     86.35    61.043    76.228
    2023-01-09    85.904    60.479    45.534
    2023-01-10     60.87    60.461    29.131
    2023-01-11    31.983     60.05    44.485
    2023-01-12    34.402      60.3     74.91
    2023-01-13    60.478    60.255    89.784
    2023-01-14    86.566     59.59    74.885
    2023-01-15    87.031    60.608    43.384
    2023-01-16    60.166    59.714      31.5

figure
plot(Data.Date, Data{:,2:end})
grid

VN = Data.Properties.VariableNames;

figure
plot(Data.Date, Data{:,2:end})
grid
xlim([datetime('1-Jun-2023') datetime('7-Jun-2023')])

Weekday = weekday(Data.Date);
WeeklyData = accumarray(Weekday, (1:numel(Weekday)).', [], @(x){Data{x,2:end}})

WeeklyData = 7×1 cell array
    {53×3 double}
    {53×3 double}
    {52×3 double}
    {52×3 double}
    {52×3 double}
    {52×3 double}
    {52×3 double}

MeanC = cellfun(@mean, WeeklyData, 'Unif',0)

MeanC = 7×1 cell array
    {[60.4105 59.9299 59.8582]}
    {[60.0404 60.1430 59.4974]}
    {[59.4016 60.0599 59.1757]}
    {[59.1616 60.0115 60.1483]}
    {[59.4423 59.8680 60.7275]}
    {[60.4756 59.6932 60.8323]}
    {[60.8717 60.0839 59.9478]}

StdeC = cellfun(@(x)std(x)/sqrt(numel(x)), WeeklyData, 'Unif',0)

StdeC = 7×1 cell array
    {[1.6846 0.0794 1.7129]}
    {[1.7274 0.0738 1.6822]}
    {[1.7232 0.0850 1.6994]}
    {[1.6920 0.0929 1.7300]}
    {[1.7262 0.0821 1.7206]}
    {[1.7278 0.0714 1.7012]}
    {[1.7138 0.0814 1.7216]}

Mean = cell2mat(MeanC)

Mean = 7×3

60.4105 59.9299 59.8582 60.0404 60.1430 59.4974 59.4016 60.0599 59.1757 59.1616 60.0115 60.1483 59.4423 59.8680 60.7275 60.4756 59.6932 60.8323 60.8717 60.0839 59.9478

Stde = cell2mat(StdeC)

Stde = 7×3

1.6846 0.0794 1.7129 1.7274 0.0738 1.6822 1.7232 0.0850 1.6994 1.6920 0.0929 1.7300 1.7262 0.0821 1.7206 1.7278 0.0714 1.7012 1.7138 0.0814 1.7216

figure
hold on
for k = 1:size(Mean,2)
    errorbar((1:7), Mean(:,k), Stde(:,k), 'DisplayName',VN{k+1}) 
end
hold off
grid
ylabel('Concentration (ng/m^3)')
xticklabels({'Sun','Mon','Tue','Wed','Thr','Fri','Sat'})
legend('Location','best')

It should easily adapt to more columns (variables) in ‘Data’.

.

[Olivia Hakan]

this worked, thanks so much!

[Star Strider]

As always, my pleasure!

Answer 2

[Image Analyst]

Look into splitapply, grpstats. and groupsummary

They will do functions, like mean, based on what group the data is in.