Eigenvalues: The Vectors That Matrices Love Most

Eigenvalues: Why Matrices Have Favorite Vectors

Every linear algebra textbook hits you with the formula Av = lambda v and expects you to be thrilled. But what does it actually MEAN?

Imagine you're stretching a photo horizontally. Most lines change direction. But some lines - the horizontal ones - don't change direction at all. They just get longer. These "direction-preserving" lines are eigenvectors, and how much they stretch is the eigenvalue.

Another example: stir water with a stick. Water swirls in all directions, but one special direction follows the stick. That's the eigenvector. How fast it flows is the eigenvalue.

Why should YOU care?

• Google's PageRank algorithm? That's just finding the principal eigenvector of a giant matrix representing link relationships between web pages. The entire Google empire started with an eigenvector.

• PCA (Principal Component Analysis) for data science? Find eigenvectors of the covariance matrix. Reduce 100 dimensions to 2 while keeping the most important patterns.

• Image compression, facial recognition, vibration analysis, quantum mechanics - eigenvectors are everywhere.

Next time someone says linear algebra is useless, hit them with PageRank. Without eigenvectors, they'd still be searching on Yahoo.