Skip to main content

🧭 Eigenvalues & Eigenvectors

These sound scary, but represent an intuitive concept: Finding the core directions of data.

🌪️ The Tornado Analogy​

Imagine a tornado hitting a house. The house gets twisted and stretched. Most of the wooden planks change direction. But maybe there is one sturdy wooden pole in the center that doesn't change its orientation—it just gets stretched taller.

  • Eigenvector: That sturdy pole. A direction that doesn't change.
  • Eigenvalue: How much the pole got stretched (e.g., if it got twice as tall, the eigenvalue is 2).

🐍 Python Implementation​

Finding eigenvectors is mathematically complex, but numpy.linalg does it in one line!

import numpy as np

# Define a transformation matrix (our "tornado")
matrix = np.array([
    [4, -2],
    [1,  1]
])

# Calculate Eigenvalues and Eigenvectors
eigenvalues, eigenvectors = np.linalg.eig(matrix)

print("Eigenvalues (How much it stretched):\n", eigenvalues)
print("Eigenvectors (The sturdy poles):\n", eigenvectors)

🎨 Visual Representation​

A glowing 3D tornado hitting a house with a solid central pillar