Preparing Features for Python

Start With Your Dataset

Imagine that we want to predict house prices.

Our dataset contains:

We already know from the previous topics that Size, Bedrooms, and Location are features.

Price is the label we want the model to predict.

Put the Features Into X

In Python, we commonly use X to represent the input features.

Think of X as:

Put the Label Into y

We commonly use y to represent the target or label.

Think of y as:

X and y Together

Now we can connect the two ideas.

During training, the model receives X together with the known values of y and learns the relationship between them.

A Simple Python Example

We can represent a simple house dataset in Python like this:

import pandas as pd

data = {
    "size": [1500, 2000, 2500],
    "bedrooms": [2, 3, 4],
    "price": [300000, 400000, 550000]
}

df = pd.DataFrame(data)

X = df[["size", "bedrooms"]]
y = df["price"]

Here, X contains the features we selected: size and bedrooms.

y contains the value we want to predict: price.

What Does X Actually Look Like?

When we select multiple features, X contains multiple columns.

X

   size    bedrooms
   1500       2
   2000       3
   2500       4

Each row represents one example, such as one house.

Each column represents one selected feature.

What Does y Look Like?

The target contains the answer for each training example.

y

300000
400000
550000

These values correspond to the houses in X.

The model learns from many of these feature-and-answer pairs.

What About Categorical Features?

Earlier we learned that features can also be categorical, such as city or membership type.

For example:

X

   size    bedrooms    city
   1500       2        City A
   2000       3        City A
   2500       4        City B

This is useful for understanding the dataset, but there is an important issue: many machine learning algorithms cannot directly work with text such as "City A" and "City B".

We will learn how to convert categorical values into a usable numerical representation later in Lesson 6 — Data Preparation.

The Important Pattern

What Goes Into X and y?

We want to predict whether a customer will purchase a product.