How to plot 2 columns for 5 data set categories with different length using boxplot

Question

bassant tolba 2023-5-12

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1962149-how-to-plot-2-columns-for-5-data-set-categories-with-different-length-using-boxplot

评论： the cyclist 2023-7-9

I want to plot the Boxplots for 2 repeated variables collected for 5 data sets, where each data set has different length (10x1, 20x1,30x1,40x1,50x1). So i actually want to plot 5 catagories on x-axis, where each catagory will have 2 vertical boxplots. matlab code

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

bassant tolba 2023-5-12

I mean simething like this, but each category has different length

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

the cyclist 2023-5-12

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1962149-how-to-plot-2-columns-for-5-data-set-categories-with-different-length-using-boxplot#answer_1233814

There is a very similar example in the documentation for the boxchart function.

To give more specific advice than that, we'd probably need to see how your data are stored. You can upload it using the paper clip icon in the INSERT section of the toolbar.

11 个评论
显示 9更早的评论隐藏 9更早的评论

the cyclist 2023-5-12

在 MATLAB Online 中打开

By far the most difficult part of this is structuring your data into the proper form to do the chart. Upstream, it would have been much better to have already structured the data better (e.g. with different names, in one MAT file, etc.)

That being said, see the code below.

(Note that there are some much more compact ways to write this code, looping over the 10-50 values, but I wanted you to see the basic idea, without needing to understand some more sophisticated methods. In particular, my using dynamic variable names here is a very bad coding practice, but I think in this case will be a better way for you to see what is going on.)

The gist is that for each file, I needed to:

Load the data into a structure
From the data, create a table that also included information about it's data type and length
Concatenate all the individual tables into one table, for plotting purposes
Plot the data

% Load the matching data into structures

S10m = load('total_matching_delay10.mat')';

S20m = load('total_matching_delay20.mat')';

S30m = load('total_matching_delay30.mat')';

S40m = load('total_matching_delay40.mat')';

S50m = load('total_matching_delay50.mat')';

% Create tables for each set of matching data

T10m = table(repmat(categorical({'Matching'}),10,1),repmat(categorical(10),10,1),S10m.dtotal_concat','VariableNames',{'Type','Length','Value'});

T20m = table(repmat(categorical({'Matching'}),20,1),repmat(categorical(20),20,1),S20m.dtotal_concat','VariableNames',{'Type','Length','Value'});

T30m = table(repmat(categorical({'Matching'}),30,1),repmat(categorical(30),30,1),S30m.dtotal_concat','VariableNames',{'Type','Length','Value'});

T40m = table(repmat(categorical({'Matching'}),40,1),repmat(categorical(40),40,1),S40m.dtotal_concat','VariableNames',{'Type','Length','Value'});

T50m = table(repmat(categorical({'Matching'}),50,1),repmat(categorical(50),50,1),S50m.dtotal_concat','VariableNames',{'Type','Length','Value'});

% Load the joint data into structures

S10j = load('total_joint_delay10.mat')';

S20j = load('total_joint_delay20.mat')';

S30j = load('total_joint_delay30.mat')';

S40j = load('total_joint_delay40.mat')';

S50j = load('total_joint_delay50.mat')';

% Create tables for each set of joint data

T10j = table(repmat(categorical({'Joint'}),10,1),repmat(categorical(10),10,1),S10j.dtotal_concat','VariableNames',{'Type','Length','Value'});

T20j = table(repmat(categorical({'Joint'}),20,1),repmat(categorical(20),20,1),S20j.dtotal_concat','VariableNames',{'Type','Length','Value'});

T30j = table(repmat(categorical({'Joint'}),30,1),repmat(categorical(30),30,1),S30j.dtotal_concat','VariableNames',{'Type','Length','Value'});

T40j = table(repmat(categorical({'Joint'}),40,1),repmat(categorical(40),40,1),S40j.dtotal_concat','VariableNames',{'Type','Length','Value'});

T50j = table(repmat(categorical({'Joint'}),50,1),repmat(categorical(50),50,1),S50j.dtotal_concat','VariableNames',{'Type','Length','Value'});

% Concatenate all the data into one table

tbl = [T10m; T20m; T30m; T40m; T50m; T10j; T20j; T30j; T40j; T50j];

% Make the box plot

figure

boxchart(tbl.Length,tbl.Value,'GroupByColor',tbl.Type)

ylabel('Length')

legend('Location','NorthWest')

bassant tolba 2023-5-13

second_paper_boxplot.png

Please kindly find the link for other dataset.

https://drive.google.com/drive/folders/1tlRxqY-e6zFuk8sOfMe493YKPq1VLmrA?usp=share_link

I used the same above code with changing the dimesnion to fit the dataset.

However, my dataset includes many outliers. Thus, in the plot, I can not see the bar as the above solved example as shown below

S10m = load('total_matching_delay2.mat')';

S20m = load('total_matching_delay4.mat')';

S30m = load('total_matching_delay6.mat')';

S40m = load('total_matching_delay8.mat')';

S50m = load('total_matching_delay10.mat')';

% Create tables for each set of matching data

T10m = table(repmat(categorical({'Matching Algorithm'}),100,1),repmat(categorical(100),100,1),S10m.dtotal_concat','VariableNames',{'Type','Length','Value'});

T20m = table(repmat(categorical({'Matching Algorithm'}),200,1),repmat(categorical(200),200,1),S20m.dtotal_concat','VariableNames',{'Type','Length','Value'});

T30m = table(repmat(categorical({'Matching Algorithm'}),300,1),repmat(categorical(300),300,1),S30m.dtotal_concat','VariableNames',{'Type','Length','Value'});

T40m = table(repmat(categorical({'Matching Algorithm'}),400,1),repmat(categorical(400),400,1),S40m.dtotal_concat','VariableNames',{'Type','Length','Value'});

T50m = table(repmat(categorical({'Matching Algorithm'}),500,1),repmat(categorical(500),500,1),S50m.dtotal_concat','VariableNames',{'Type','Length','Value'});

% Load the joint data into structures

S10j = load('total_joint_delay2.mat')';

S20j = load('total_joint_delay4.mat')';

S30j = load('total_joint_delay6.mat')';

S40j = load('total_joint_delay8.mat')';

S50j = load('total_joint_delay10.mat')';

% Create tables for each set of joint data

T10j = table(repmat(categorical({'Joint Algorithm'}),100,1),repmat(categorical(100),100,1),S10j.dtotal_concat','VariableNames',{'Type','Length','Value'});

T20j = table(repmat(categorical({'Joint Algorithm'}),200,1),repmat(categorical(200),200,1),S20j.dtotal_concat','VariableNames',{'Type','Length','Value'});

T30j = table(repmat(categorical({'Joint Algorithm'}),300,1),repmat(categorical(300),300,1),S30j.dtotal_concat','VariableNames',{'Type','Length','Value'});

T40j = table(repmat(categorical({'Joint Algorithm'}),400,1),repmat(categorical(400),400,1),S40j.dtotal_concat','VariableNames',{'Type','Length','Value'});

T50j = table(repmat(categorical({'Joint Algorithm'}),500,1),repmat(categorical(500),500,1),S50j.dtotal_concat','VariableNames',{'Type','Length','Value'});

% Concatenate all the data into one table

tbl = [T10m; T20m; T30m; T40m; T50m; T10j; T20j; T30j; T40j; T50j];

% Make the box plot

figure

boxchart(tbl.Length,tbl.Value,'GroupByColor',tbl.Type);

xlabel('Number of TNs')

ylabel('Total sum delay')

legend('Location','NorthWest')

k;

Please can you help ? Thank you.

the cyclist 2023-5-13

在 MATLAB Online 中打开

I mentioned that there was a more compact way of writing this code (rather than using dynamically named variables). It's also more robust against typos and other errors.

It could be made even more "automatic" by detecting the names of the files in the directory.

% Initialize empty table
var_names_types = [["Type", "categorical"]; ...
			       ["Length", "categorical"]; ...
			       ["Value", "double"]; ...
                  ];
% Make table using fieldnames & value types from above
tbl = table('Size',[0,size(var_names_types,1)],... 
	'VariableNames', var_names_types(:,1),...
	'VariableTypes', var_names_types(:,2));
% Loop over the data files
for n = [2 4 6 8 10]
    % Load "matching" data into table
    S = load(sprintf('total_matching_delay%d.mat',n));
    TM = table(repmat({'Matching'},numel(S.dtotal_concat),1) ,repmat(categorical(n), numel(S.dtotal_concat),1), S.dtotal_concat', 'VariableNames',{'Type','Length','Value'});
    % Load "joint" data into table
    S = load(sprintf('total_joint_delay%d.mat',n));
    TJ = table(repmat({'Joint'},   numel(S.dtotal_concat),1) ,repmat(categorical(n), numel(S.dtotal_concat),1), S.dtotal_concat', 'VariableNames',{'Type','Length','Value'});
    % Append new data to existing table
    tbl = [tbl; TM; TJ];
end
% Make the box plot
figure
boxchart(tbl.Length,tbl.Value,'GroupByColor',tbl.Type)
ylabel('Length')
legend('Location','NorthWest')
% % Uncomment this line to adjust the y limit, if needed
% set(gca,"YLim",[0 5])

bassant tolba 2023-7-9

Please do you know how to change the color of one if those two boxbar plots???

Thank you so much

the cyclist 2023-7-9

在 MATLAB Online 中打开

Set the BoxFaceColor and MarkerColor property of that boxchart object.

Here is an example, based on the documentation:

tbl = readtable('TemperatureData.csv');
monthOrder = {'January','February','March','April','May','June','July', ...
    'August','September','October','November','December'};
tbl.Month = categorical(tbl.Month,monthOrder);
figure
hb = boxchart(tbl.Month,tbl.TemperatureF,'GroupByColor',tbl.Year); % <--- Note that I assigned the handle hb to the boxchart
ylabel('Temperature (F)')
legend
% Pick a color. (Here I used an RGB value, but there are other options. See
% documentation.)
newColor = [0 1 1];
% Set the box and marker color to the new color
hb(2).BoxFaceColor = newColor;
hb(2).MarkerColor  = newColor;