0 of 6 done
Start hereEigenvalues and Eigenvectors

The mathematics of things that add and scale. Written to the course standard: prerequisites, observable outcomes, worked derivations, a runnable lab whose output was generated from an actual run, practice with hidden answers, and a demonstrable finish line.

This umbrella merges both curricula โ€” SS3 matrices, determinants and vectors from the NERDC outline, and the Year 1โ€“2 university track โ€” so each subject appears once, secondary treatment first.

The one idea: a matrix is not a table of numbers, it is a function. Everything else follows. The multiplication rule is function composition; the determinant is the area it scales by; the column space is what it can output; the null space is what it destroys; and eigenvectors are the directions it does not turn. Learning the table without the function is why linear algebra feels arbitrary.

Reading order

Dependencies are strict. Matrices before systems, systems before vector spaces, vector spaces before eigenvalues.

  1. Matrices and Determinants โ€” [Intermediate] โ€” the operations, and what a determinant actually measures: the factor by which area is scaled
  2. Vectors โ€” [Intermediate] โ€” the algebra that works identically in 2 or 50,000 dimensions, Cauchyโ€“Schwarz derived, and why cosine similarity compares documents
  3. Systems of Linear Equations โ€” [Intermediate] โ€” how software really solves , and the one-line change without which it returns wrong answers silently
  4. Vector Spaces โ€” [Advanced] โ€” the abstraction that makes polynomials vectors, and the rankโ€“nullity theorem
  5. Linear Transformations โ€” [Advanced] โ€” why the multiplication rule is what it is, derived rather than asserted
  6. Eigenvalues and Eigenvectors โ€” [Advanced] โ€” the directions a map leaves alone, diagonalisation, and PageRank in miniature

The thread through the course

Three questions are asked in lesson 1 and answered in lesson 5:

AskedAnswered
Why is multiplication row-times-column rather than elementwise?It is function composition, written down โ€” [[05-linear-transformations/01-linear-transformations
Why is ?Because โ€œrotate then scaleโ€ is not โ€œscale then rotateโ€
Why does mean no inverse?The map crushes area to zero, so information is destroyed and cannot be recovered

And two more run the length of the course: rank appears as a pivot count in lesson 3, as a dimension in lesson 4, and as the size of an image in lesson 5 โ€” the same number, three descriptions. The null space is a solution set in lesson 3, a subspace in lesson 4, a kernel in lesson 5, and an eigenspace in lesson 6.

Where the two curricula meet

LessonSecondary sourceUniversity source
Matrices and determinantsSS3 T1 โ€” operations, determinants, inversesY1S2 โ€” general matrices, Cramerโ€™s rule
VectorsSS3 T1 โ€” components, magnitude, scalar productY1S2 โ€” vector algebra, dimensions
Systems of linear equationsโ€”Y1S2 โ€” Gaussian elimination, inverse matrices
Vector spacesโ€”Y2S1 โ€” subspaces, independence, bases, dimension
Linear transformationsโ€”Y2S1 โ€” kernel and image
Eigenvaluesโ€”Y2S1 โ€” eigenvalues and eigenvectors

What is verified

Every lab was executed and its โ€œExpected outputโ€ block generated from that run. Several check a claim the prose makes rather than illustrating it:

  • The determinant is an area factor โ€” six transformations applied to the unit square, and the image area equals every time, including a shear (area unchanged) and a singular matrix (area zero).
  • Pivoting is not optional โ€” the same system solved both ways: without pivoting , with pivoting . No error, no warning, 100% wrong.
  • Rank + nullity = columns โ€” verified across five matrices including a of zeros.
  • Differentiation is a matrix โ€” polynomials as a vector space, with โ€œthe only functions with zero derivative are the constantsโ€ computed as a null space.
  • Rotation matrices multiply by adding angles โ€” which is the trigonometric addition formula, recovered independently of its derivation in trigonometry.
  • A Markov chain forgets where it started โ€” two opposite initial states, both landing on to nine decimal places.

How to study this

  1. Attempt the prediction before reading on. Every lesson stops once and asks you to commit. Answers are at the end of Check your understanding, never beside the question.
  2. Hold the function picture from the start. If a rule seems arbitrary, ask what it means for the map rather than for the table.
  3. Do the labs and break them. Each is standard library only, self-contained, and writes no files. The pivoting lab in particular is worth running with your own matrices.
  4. Watch for exact versus floating-point arithmetic. Lessons 4 and 5 use Fraction deliberately, because โ€œis this pivot zero?โ€ cannot be answered honestly in floats โ€” a point lesson 3 demonstrates the hard way.
โ†‘ Back to Mathematics