Week: 11 Debugging Journey- Fixing the tukey_multiple Function

Introduction

In this post, I’ll explain how I fixed a bug in the tukey_multiple function in R. The goal of this function is to find rows in a dataset where all the values are outliers. Initially, I ran into errors, but I eventually found the problem and fixed it. Here’s how I did it.

Understanding the tukey_multiple Function

The function tukey_multiple is supposed to:

1. Check each column in a dataset to see if values are outliers.
2. Return a list that tells us whether each row has only outliers across all columns.

Here’s the original code I started with:

Step 1: Initial Problem - Missing Function

When I tried to run the code, I got an error saying tukey.outlier wasn’t found. After checking, I realized that tukey.outlier isn’t a built-in function in R and wasn’t defined anywhere else in the code. This function is supposed to check if values in a column are outliers, so I knew I needed to write my own function to do that.

What I Learned

• Check if functions exist before using them, especially if they’re not from R’s built-in functions or packages.

Step 2: Creating a Helper Function for Outliers

To fix the missing function, I created my own tukey_outlier function. This function detects outliers based on Tukey’s Rule, which states that:

• Values below Q1 - 1.5 * IQR or above Q3 + 1.5 * IQR are outliers.
• Here, Q1 is the first quartile (25th percentile), Q3 is the third quartile (75th percentile), and IQR is the interquartile range (Q3 - Q1).

Here’s my new helper function:

What I Learned

• Understanding Tukey’s Rule helped me create a simple function to detect outliers without needing a special library.

Step 3: Fixing the Logical Error

Once I defined tukey_outlier, I ran into a new issue with this line of code:

The && operator didn’t work correctly here because && is used for single logical comparisons in R, not for checking every element in an array. So, I replaced && with a simpler check that applied tukey_outlier to each column directly:

What I Learned

• Use & (element-wise comparison) instead of && when working with arrays in R.

Step 4: Improving Row-Wise Check for Outliers

Lastly, I wanted to check if each row had only outliers across all columns. Originally, I used a loop to check each row

To make this part simpler, I replaced the loop with apply(outliers, 1, all). This single line does the same thing, checking if every value in each row is an outlier

Final Solution

Here’s the final, corrected tukey_mutiple function:

Testing and Results

I tested the fixed function on the iris dataset, and it worked as expected. Running tukey_multiple(iris) gave me a vector that showed which rows had all outliers.

Conclusion

Through this debugging process, I learned:

1. The importance of defining all functions used in code.
2. How to apply Tukey’s Rule for finding outliers.
3. The difference between & and && for logical operations.
4. How to simplify code with apply for row-wise or column-wise operations.

This experience helped me understand debugging better and taught me some useful R tricks for handling data more efficiently.

For a more detailed exploration of the data manipulation techniques and code examples, you can find everything in my GitHub repository.