Plotting categorical array of varying size

Hi all,

I have 3 array types I'm using to create a bar graph:

• (variable c) a categorical array with character vectors where a type of category might exist more than once [created using built-in function categorical]
• (variable x) a cell array containing the different types of vectors that exist in variable c [created using built-in function categories]
• (variable xx) a double array, containing the number of occurrances variable c exists for each type of variable x [created using countcats]

I'm using bar(xx), and pareto(xx) to plot this data. However, I would like to plot this data from a folder of 1000+ txt files into a single figure, where each file doesn't necessarily contain common variable x. Therefore as I loop through each txt file, the categorical array rows ≠ columns.

Should I make this a cell array to manipulate here, or maybe use addcats function? I've written a function that works correctly for a single txt file, but I am unsure how to integrate these 3 array types to where I can plot aggregate data from all txt files.

Please let me know additional code you think may be helpful -- didn't seem useful to begin with.

Thanks,

Jarrod

Answer 1

[Shlok]

Hi Jarrod,

As per my understanding, you have multiple files containing categorical arrays, and you want to read through them and plot a histogram.

To do this, I suggest reading all the files, concatenating the categorical values into a global array, and then generating a histogram by extracting unique categories and their counts using the “categories” and “countcats” functions.

Here’s how you can achieve that:

1. Start by initializing a categorical array that will store data from all the files.

allData = categorical([]); 

2. Now, loop through each file, read the data and convert the relevant column or portion of the data to a categorical array.

% 'data' is the output of readtable, which reads the file within the loop 
curFileData = categorical(data.YourColumnName);  

3. After reading the data from each file, concatenate the current categorical array with the main array.

allData = [allData; curFileData]; 

4. Use the “categories” function to find the unique categories in the concatenated array.

uniqueCategories = categories(allData); 

5. Use the “countcats” function to get the count of each category.

counts = countcats(allData); 

6. Finally, plot a bar graph using the counts for each category.

bar(counts); 
xticklabels(uniqueCategories); 

Refer to the following MathWorks Documentation link, to know more about “countcats” function:

https://www.mathworks.com/help/releases/R2020b/matlab/ref/categorical.countcats.html

Hope this helps.