In this recipe, it is very important to create categorical columns, as they are allowed to be ordered. Seaborn uses this ordering to place the labels on the plot. Steps 3 and 4 show what clearly appears to be a downward trend for increasing diamond quality. This is where Simpson's paradox takes center stage. This aggregated result of the whole is being confounded by other variables not yet examined.

The key to uncovering this paradox is to focus on carat size. Step 5 reveals to us that carat size is also decreasing with increasing quality. To account for this fact, we cut the diamond size into five equally-sized bins with the qcut function. By default, this function cuts the variable into discrete categories based on the given quantiles. By passing it an integer, as was done in this step, it creates equally-spaced quantiles. You also have the option of passing it a sequence of explicit non-regular quantiles.

With this new variable, we can make a plot of the mean price per diamond size per group, as done in step 6. The point plot in seaborn creates a line plot connecting the means of each category. The vertical bar at each point is the standard deviation for that group. This plot confirms that diamonds do indeed become more expensive as their quality increases, as long as we hold the carat size as the constant.