Imagine you are working on an assembly line, where each worker performs a specific task before passing the product to the next person. This setup is efficient because each action builds upon the previous one, and the process flows smoothly from start to finish. In data analysis, you often need to perform a series of transformations on your data—such as selecting columns, filtering rows, and arranging results. Chaining these operations together can make your workflow more efficient and your code easier to follow, just like a well-organized assembly line.

In the pipeline above, you start with the iris dataset. First, you use select() to keep only the Species, Sepal.Length, and Sepal.Width columns. Next, you apply filter() to retain only rows where Sepal.Length is greater than 5. Finally, you use arrange() to sort the results in descending order by Sepal.Width. Each function takes the output of the previous step as its input, allowing you to build a sequence of transformations that is both logical and easy to follow.

When you compare the two approaches above, it is clear that using pipes makes your code more readable and maintainable. Without pipes, you have to read the code inside out, starting from the innermost function and working your way outward. This can become confusing, especially as the number of steps increases. With pipes, you read the code from top to bottom, in the same order as the transformations are applied. This makes it easier to understand what is happening at each stage and to make changes or debug your workflow in the future.