Encoding Categorical Data

Machine learning models work with numbers, not text, so you must convert categorical data like colours or city names into numerical form before training.

Common encoding methods

• Label Encoding, which assigns a number to each category.
• One-hot encoding, which creates separate columns for each category.

Label Encoding

Useful when categories have a natural order, like low, medium, and high.

from sklearn.preprocessing import LabelEncoder

encoder = LabelEncoder()

df["Size"] = encoder.fit_transform(df["Size"])

One-Hot Encoding

Useful when categories have no order, like city names or colors.

df = pd.get_dummies(df, columns=["City"])

get_dummies() creates a new column for each unique category, marking it as 1 or 0 based on presence, which avoids implying any false ranking between categories.

The right encoding method depends on whether your categories have a meaningful order.