Heatmaps for Correlation

Correlation recap

Correlation describes how two numeric variables move together.

• +1: strong positive relationship
• -1: strong negative relationship
• 0: no linear relationship

diagram From a DataFrame to a heatmap mermaid

  flowchart LR
  A["Numeric columns"] --> B["df.corr()
(square correlation matrix)"]
  B --> C["sns.heatmap(corr)"]
  C --> D["Color grid
(+ optional annot=True numbers)"]

A correlation matrix is a square grid of numbers, which sns.heatmap turns into a grid of colors.

Caution

A heatmap’s color scale is only as trustworthy as the correlation numbers behind it. Always fix vmin=-1, vmax=1 (as in the example below) so the color scale is comparable across different plots — otherwise a weak correlation can look just as “hot” as a strong one if the colormap auto-scales to a narrower range.

Correlation heatmap

Correlation heatmap

import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
 
penguins = sns.load_dataset("penguins").dropna()
 
corr = penguins[[
    "bill_length_mm",
    "bill_depth_mm",
    "flipper_length_mm",
    "body_mass_g",
]].corr()
 
plt.figure(figsize=(6, 4))
sns.heatmap(corr, annot=True, cmap="coolwarm", vmin=-1, vmax=1)
plt.title("Penguins correlation")
plt.tight_layout()
plt.show()

Visualize it

Each cell’s color encodes a correlation value from -1 to +1 — amber for strong positive, blue for strong negative, and dim in the middle for near zero:

sketch Correlation matrix as a color grid p5.js Pause Restart

Each cell's shade encodes its correlation value: amber for positive, blue for negative, dim for near zero.

Tips

• Correlation is not causation.
• Outliers can inflate correlations.
• For non-linear relationships, correlation can be misleading.

Next

Continue to Pair Plots to see the raw scatter relationships behind each correlation number.

🧪 Try It Yourself

Exercise 1 – Compute a Correlation Matrix

pch.exerciseTag pch.exerciseTitle

Exercise 2 – Draw the Heatmap With Numbers

pch.exerciseTag pch.exerciseTitle

Exercise 3 – Fix the Color Scale

pch.exerciseTag pch.exerciseTitle