Change of Basis

Same vector, different coordinates. Learn why choosing the right basis makes everything simpler.

Change of Basis

A vector is a geometric object—an arrow in space. But the numbers we use to describe it depend on our choice of basis.

Changing basis is like changing languages: the same object, different descriptions. The right basis can make complicated problems simple.

The Core Idea

Every vector can be written as a linear combination of basis vectors.

v = c₁b₁ + c₂b₂ + ... + cₙbₙ

Different bases → different coefficients (c₁, c₂, ...) → same vector!

Demo 1: Same Vector, Different Coordinates

Drag the point and watch its coordinates change in both the standard basis (blue/green) and a custom basis (purple/pink). Same location, different numbers!

Standard Basis
x = 2.00,y = 1.50
v = 2.00ê₁ + 1.50ê₂
Custom Basis
c₁ = 1.54,c₂ = 0.61
v = 1.54b₁ + 0.61b₂

Drag the orange point. The same point in space has different coordinates depending on which basis you use. The point doesn't move—only how we describe it changes.

Demo 2: The Change of Basis Matrix

To convert coordinates between bases, we use a matrix. P⁻¹ converts from standard to new basis; P converts back.

[v]B=P⁻¹·[v]std
P⁻¹ converts from standard to new basis

The vector v in standard coordinates: (2, 1)

Standard Coords
(2, 1)
New Basis Coords
(1.38, 0.26)
P (basis matrix)
[1.5, -0.3]
[0.5, 1.2]
P⁻¹ (inverse)
[0.62, 0.15]
[-0.26, 0.77]

Demo 3: Why Change Basis?

In the standard basis, a transformation might look complicated. But in the eigenbasis, the same transformation becomes pure scaling—no shearing!

Matrix A (standard)
[2, 1]
[1, 2]
Off-diagonal terms cause shearing
Matrix D (eigenbasis)
[3.0, 0]
[0, 1.0]
Pure scaling—no shearing!

Key insight: The same transformation looks complicated in the standard basis but becomes pure scaling in the eigenbasis. The off-diagonal terms disappear!

In the eigenbasis, each axis just stretches by its eigenvalue (λ₁ = 3.0, λ₂ = 1.0). This is why eigenvectors are the "natural" coordinates for understanding transformations.

Demo 4: Change of Basis in 3D

The same ideas extend to 3D. Toggle between standard and custom bases to see how the same point gets different coordinates.

Standard Coords
x=1.50, y=1.00, z=0.80
Custom Coords
c₁=1.32, c₂=0.71, c₃=0.53

The white point stays in the same location in space, but its coordinates change depending on which basis you use. Drag to rotate the view and see both coordinate systems.

Key Takeaways

  • Coordinates depend on basis: The same vector has different numbers in different bases
  • P matrix: Columns are the new basis vectors expressed in standard coordinates
  • P⁻¹ converts: [v]new = P⁻¹ · [v]std
  • Eigenbasis simplifies: Transformations become diagonal (pure scaling) in the eigenbasis
  • Similarity: A and P⁻¹AP represent the same transformation in different bases

Choosing the right basis is one of the most powerful techniques in linear algebra!