Categories as Numbers

Models do math, so this chapter turns words into numbers. Ordered categories, like juice sizes, get numbers in their order:

A dictionary holds the numbers; .map looks up every value, and square brackets look up one:

orders = pd.DataFrame({"juice": ["mango", "lime", "berry", "orange"],
                       "size": ["large", "small", "medium", "small"]})
sizes = {"small": 1, "medium": 2, "large": 3}
orders["size_code"] = orders["size"].map(sizes)
print(orders)
print(sizes["medium"])

This prints:

    juice    size  size_code
0   mango   large          3
1    lime   small          1
2   berry  medium          2
3  orange   small          1
2

As text the order is lost: "large" < "small" is True, alphabetically. The codes keep it.

A word missing from the dictionary, like "Large" with a capital L, becomes NaN: count those with .isna().sum().