SpecEnvoy
EN·中文

Matrix Transformations

The idea that unlocks all of linear algebra: a matrix is not a table of numbers, it's an instruction for bending space. Every point in the plane moves — but in a way so orderly that just two arrows tell you everything. First-year university.

The columns are where the basis lands

Start with the two unit arrows: î pointing right, ĵ pointing up. A 2×2 matrix does one thing — it says where those two arrows should go. Its first column is the new î, its second column is the new ĵ:

M = [abcd]  →  î lands at (a, c),   ĵ lands at (b, d)

Everything else follows for free. Because the transformation is linear — grid lines stay straight, parallel and evenly spaced — once you've pinned down î and ĵ, every other point is just a combination of the two. Move the two arrows and the entire plane comes along.

Bend the plane 🟠🔵

Drag the orange î and blue ĵ tips (they snap to half-units), or use the sliders and preset buttons. Watch the green unit square — its area is the determinant.

The determinant is an area

The green box is the image of the unit square. Its area is exactly |det M| = |ad − bc| — the factor by which the transformation scales every area in the plane. That one number carries a lot:

Zero → collapse

When det = 0 the box flattens to a line: the whole plane gets squashed onto a lower dimension. Information is lost, and the matrix has no inverse — you can't undo a squash. Try the "Squash" preset.

Negative → flipped

A negative determinant means the transformation turned the plane over, like a page — orientation reversed. î and ĵ have swapped their clockwise/counter-clockwise order. Try "Reflect".

So det = 2 doubles areas, det = 1 (rotations, shears) preserves them, det = 0.5 halves them, and a negative det does all that and mirrors. It's the single most useful number a matrix carries.

A matrix is the many-dimensional cousin of the Function Transformer's stretch-and-slide. The determinant's "area of a parallelogram" is the same quantity the cross product computes in 3D. Back to Linear algebra.