How to generate boxplot from data in struct

Let's say I have a struct that contains a bunch of arrays of data of different sizes. What is the easiest way to generate a boxplot from this data?

Answer 1

[Jatin Singh]

Hi Asef Islam,

To do this, we must first format the data which can be read by the “boxplot” function. Let’s the group the data in each field using their field name, this can be done by combining all array into a single column vector.

For e.g., let's say the struct is “dataStruct” defined as:

%initialize dataStruct
dataStruct.field1 = rand(10, 1);  
dataStruct.field2 = rand(15, 1); 
dataStruct.field3 = rand(12, 1); 

Using the following code boxplot can be generated from a struct with arrays as fields.

% Initialize empty arrays for data and group labels 
data = []; 
labels = [];

% Get the field names of the struct 
fields = fieldnames(dataStruct); 

% Loop through each field in the struct
for i = 1:length(fields) 
     % Extract the data array from the struct 
     cdata = dataStruct.(fields{i}); 
     % Concatenate the array with data 
     data = [data; cdata]; 
     % Create a group label array for the current data 
     labels = [labels; repmat({fields{i}}, length(cdata), 1)];
end

% Generate a boxplot  
boxplot(data, labels); 

% Adding title and labels 
title('Boxplot of Struct Data'); 
xlabel('Data Fields'); 
ylabel('Values'); 

You can also refer to the MathWorks documentation of “boxplot” for more details:

https://www.mathworks.com/help/stats/boxplot.html

Hope this helps.