If you like proofs and intuition
Linear Algebra: Theory, Intuition, Code
Strong focus on proofs, with coding and applications.
Linear algebra
The language of data science, machine learning, and signal processing.
Regression, PCA, neural networks, source separation: almost every method in modern data analysis is linear algebra underneath. I've written two books and a 30-hour course on it, each with a different emphasis. Pick the one that matches how you like to learn, and use the others as companions.
Choose your starting point
Different ways into the same material. Any one of them is a complete start.
If you want to code and apply it
Practical Linear Algebra for Data Science
Strong focus on coding and applications, with proofs.
If you learn best from video
Linear algebra: theory and code
Vectors, matrix multiplication, least-squares projections, eigendecomposition, and singular value decomposition, explained clearly and implemented in code.
Everything on linear algebra
Courses, books, code, and free material in one list.
Linear algebra: theory and code
Vectors, matrix multiplication, least-squares projections, eigendecomposition, and singular value decomposition, explained clearly and implemented in code.
Linear Algebra: Theory, Intuition, Code
Strong focus on proofs, with coding and applications.
Practical Linear Algebra for Data Science
Strong focus on coding and applications, with proofs.
Linear algebra exercise solutions on video
Video walkthroughs of the exercises in Practical Linear Algebra for Data Science.
PCA and multivariate signal processing
Theory and implementation of dimension-reduction and source-separation methods for multivariate data, with sample data and a 3-hour linear algebra crash course.
Linear algebra and hidden spaces
PCA, generalized eigendecomposition, ICA, factor analysis, hidden Markov models, and linear dynamical systems, with projects on real data.
All the linear algebra code
Companion code for the courses and books on this page, ready to run and modify.
Nothing in this format yet.
Where linear algebra fits
← Comes before
Scientific coding
New to Python? Master Python for scientific coding, or Master math by learning Python, will get you ready.
Leads to →
Signal processing & neural time series
PCA, source separation, and Fourier methods put eigendecomposition and SVD to work on real multichannel signals.
Leads to →
Deep learning & LLMs
Every layer of a neural network is a matrix multiplication. See it happen inside real models with PyTorch.