Skip to content

Cheatsheet: Linear transformations as moves

A transformation is a function from vectors to vectors (picture every point moving at once). It is linear when both hold:

RequirementMeaning
Origin stays fixedThe zero vector maps to the zero vector.
Lines stay straightGrid lines stay parallel and evenly spaced; no curving, no bunching.

Allowed: stretch, rotate, shear, flip. Not allowed: curve, bend, shift the whole plane.

A linear transformation is fully determined by where it sends i-hat and j-hat.

v = x · i-hat + y · j-hat ===> L(v) = x · L(i-hat) + y · L(j-hat)

Same scalars x, y (the coordinates of v); only the two vectors they scale have changed. Track two landing spots, reconstruct every vector.

A matrix is the record of where the basis landed

Section titled “A matrix is the record of where the basis landed”
[ a b ] first column [a, c] = L(i-hat)
[ c d ] second column [b, d] = L(j-hat)

Matrix-vector product (apply the transformation):

M · v = x · (first column) + y · (second column)

This is the linear combination from Lesson 2, with the transformed basis vectors as ingredients. Not a memorized rule.

TransformationMatrixi-hat landsj-hat lands[3, 4] becomes
Horizontal stretch x2[[2, 0], [0, 1]][2, 0][0, 1][6, 4]
90 deg counterclockwise[[0, -1], [1, 0]][0, 1][-1, 0][-4, 3]

Stretch: 3·[2,0] + 4·[0,1] = [6, 4]. Rotation: 3·[0,1] + 4·[-1,0] = [-4, 3].

  1. Read the two columns: they are where i-hat and j-hat land.
  2. Plot those two image vectors from the origin.
  3. Draw the parallelogram they span. That is the image of the unit square.

Square stays a square (rotation), becomes a rectangle (axis stretch), becomes a slanted parallelogram (shear), or squashes nearly flat (columns nearly parallel, plane collapsing toward a line).

  • Reading rows instead of columns. Meaning lives in the columns: two destination vectors side by side.
  • Forgetting the origin is pinned. A fixed shift of the whole plane is affine, not linear; a 2x2 matrix cannot do it.
  • Letting the grid curve. If lines bend, it is not a linear transformation.
  • Memorizing M·v. It is just x·(col 1) + y·(col 2).
  • Linear transformation: a vector-to-vector function fixing the origin and keeping lines straight, parallel, evenly spaced.
  • Matrix: the two transformed basis vectors written as columns.
  • Matrix-vector product: reassembling a vector from the transformed basis, x·col1 + y·col2.
  • Unit square: corners at origin, i-hat, j-hat, i-hat + j-hat; its image is the parallelogram of the columns.

A matrix is not a grid of numbers; it is a record of where the basis vectors land, and that record moves all of space.