Lecture 8: Transformations

What matrices do to space — the geometry of linear maps.

The Hook

Unit Square Under Transformation

Determinant
1.00
Area Factor
1.00
Orientation
+

The matrix doesn't just transform vectors — it transforms all of space.

Understanding transformations means understanding what gets stretched, compressed, rotated, reflected, or collapsed.

Recognition

ARRANGEMENT: "Order matters."

A matrix encodes a transformation — a systematic rearrangement of space. Every point moves according to the same linear rule.

Linear Transformations

Definition: A function T: ℝⁿ → ℝᵐ is a linear transformation if:

  1. Additivity: T(u + v) = T(u) + T(v)
  2. Homogeneity: T(cv) = cT(v)

Key insight: The columns of A are the images of the standard basis vectors.

Column 1 = where (1,0) goes. Column 2 = where (0,1) goes.

Geometry of 2D Transformations

Transformation Catalog

45°

Matrix:

Standard Transformations

TransformationMatrixdet
Rotation by θ[[cos θ, -sin θ], [sin θ, cos θ]]1
Reflection across x[[1, 0], [0, -1]]-1
Reflection across y[[-1, 0], [0, 1]]-1
Scale by (sₓ, sᵧ)[[sₓ, 0], [0, sᵧ]]sₓsᵧ
Horizontal shear[[1, k], [0, 1]]1
Project onto x[[1, 0], [0, 0]]0

The Determinant

Geometric meaning: The determinant measures how a transformation scales area (2D) or volume (3D).

Determinant = Signed Area

1.5
0.5
0.3
1.2

det([[a,b],[c,d]]) = ad - bc = 1.65

Determinant Formulas

2×2:

$$\det\begin{pmatrix} a & b \\ c & d \end{pmatrix} = ad - bc$$

3×3 (cofactor expansion):

$$\det\begin{pmatrix} a & b & c \\ d & e & f \\ g & h & i \end{pmatrix} = a(ei - fh) - b(di - fg) + c(dh - eg)$$

Kernel and Image

Kernel and Image Visualizer

Kernel (null space)
{0}
Image (range)
ℝ²
Rank + Nullity
2 + 0 = 2

Kernel: ker(A) = {v : Av = 0} — vectors that get collapsed to zero

Image: im(A) = {Av : v ∈ ℝⁿ} — all possible outputs

Rank-Nullity Theorem: dim(ker) + dim(im) = n

Composition of Transformations

Order Matters: AB ≠ BA

BA =

Warning: The order is reversed!

Chemistry Connection: Symmetry Operations

Molecular Symmetry Operations

Matrix:

det = ()

OperationdetType
Identity E+1Proper
Rotation Cₙ+1Proper
Reflection σ-1Improper
Inversion i-1Improper
Improper rotation Sₙ-1Improper

Summary: Transformation Toolkit

FeatureHow to check
Preserves area?|det| = 1
Invertible?det ≠ 0
Preserves orientation?det > 0
Orthogonal?AᵀA = I
Pure rotation?Orthogonal and det = +1
Has collapse?det = 0 or nullity > 0

Exercises

Exercise 1: Identifying Transformations

For each matrix, identify the geometric transformation:

(a) [[0, 1], [-1, 0]]   (b) [[1, 0], [0, -1]]   (c) [[3, 0], [0, 3]]   (d) [[1, 2], [0, 1]]

(a) Rotation by -90° (or 270°). Check: (1,0) → (0,-1)

(b) Reflection across x-axis

(c) Uniform scaling by factor 3

(d) Horizontal shear with k = 2

Exercise 2: Determinants

Compute det and interpret geometrically:

(a) [[2, 0], [0, 3]]   (b) [[1, 1], [2, 2]]   (c) [[cos θ, -sin θ], [sin θ, cos θ]]

(a) det = 6. Area scales by 6. Stretches 2× horizontally, 3× vertically.

(b) det = 0. Singular — collapses space. Columns are parallel.

(c) det = cos²θ + sin²θ = 1. Area preserved — it's a rotation.

Exercise 3: Composition

(a) Find the matrix for: reflect across x-axis, then rotate 90°.

(b) Find the matrix for: rotate 90°, then reflect across x-axis.

(c) Are they the same?

(a) R₉₀ × Reflₓ = [[0,-1],[1,0]][[1,0],[0,-1]] = [[0,1],[1,0]]

(b) Reflₓ × R₉₀ = [[1,0],[0,-1]][[0,-1],[1,0]] = [[0,-1],[-1,0]]

(c) Different! (a) is reflection across y=x; (b) is reflection across y=-x.

Exercise 4: Symmetry Operations

(a) Write the matrix for C₃ (120° rotation about z).

(b) Verify C₃³ = E.

(c) What is det(C₃)? Proper or improper?

(a) C₃ = [[cos120°, -sin120°], [sin120°, cos120°]] = [[-1/2, -√3/2], [√3/2, -1/2]]

(b) C₃³ = rotation by 360° = I ✓

(c) det(C₃) = 1. Proper rotation (preserves orientation).