Impara Correlation Heatmaps for Feature Relationships | Bivariate and Correlation Analysis

Understanding how features relate to each other is essential in retail analytics, where many variables—such as sales, inventory, discounts, and customer behavior—can be connected. A correlation matrix lets you quickly see the strength and direction of relationships between every pair of numerical features in your dataset.

In a retail context, a correlation matrix can help you:

Identify if higher discounts are linked to increased sales;
Discover whether certain product categories tend to move together;
Spot connections between inventory levels and customer activity;
Reveal hidden patterns that influence business outcomes.

By summarizing all pairwise correlations in a single table, you get a clear overview of how your features interact. This insight can guide deeper analysis and support smarter business decisions.


              1234567891011121314
            
import pandas as pd

# Example retail dataset
data = {
    "sales": [200, 260, 240, 310, 280],
    "discount": [12, 25, 18, 10, 22],
    "inventory": [85, 95, 80, 90, 70],
    "customer_count": [40, 32, 48, 36, 44]
}
df = pd.DataFrame(data)

# Compute the correlation matrix for selected numerical features
corr_matrix = df[["sales", "discount", "inventory", "customer_count"]].corr()
print(corr_matrix)


              12345678
            
import seaborn as sns
import matplotlib.pyplot as plt

# Visualize the correlation matrix as a heatmap
plt.figure(figsize=(6, 4))
sns.heatmap(corr_matrix, annot=True, cmap="coolwarm", fmt=".2f")
plt.title("Correlation Heatmap of Retail Features")
plt.show()

When reading a correlation heatmap, focus on the following points:

Values close to 1: strong positive relationship; as one feature increases, the other tends to increase as well. For example, if sales and customer_count have a value near 1, higher customer count is associated with higher sales;
Values close to -1: strong negative relationship; as one feature increases, the other tends to decrease. For instance, a value near -1 between discount and inventory suggests higher discounts are linked to lower inventory levels;
Values close to 0: weak or no linear relationship between the features.

Look for the darkest or brightest colors in the heatmap to quickly spot feature pairs with the strongest relationships. These pairs are often the most important for deeper analysis in your retail dataset.

Tutto è chiaro?

Grazie per i tuoi commenti!

Sezione 3. Capitolo 5

Chieda ad AI

Chieda pure quello che desidera o provi una delle domande suggerite per iniziare la nostra conversazione

Suggested prompts:

Can you explain how to interpret the correlation values in the matrix?

What should I do if I find two features with a very high correlation?

How can I use these insights to improve my retail business decisions?

Awesome!

Completion rate improved to 5.56

Scorri per mostrare il menu