Typorithm learning path
NumPy: arrays and numerical computing
Build trustworthy numerical workflows through shapes, dtypes and NumPy 2 promotion, slicing, masks, ownership, broadcasting, reductions, missing values, safe ufuncs, aligned sorting/grouping, matrix products, owned random generators, verified numerical files, bounded workspace and reference-only scaling. 18 topics with six substantive sections, controlled mistakes, small tasks and two hints. On a computer, type short code and reconstruct it from memory; the app offers reading and explained questions. Complete examples execute in the downloadable local lab, verified with Python 3.13 and NumPy 2.5.4. Synthetic owned data only.
Study with a practical destination
Use the linked lessons in order, complete the lab and explain the decisions without the reference. Select the path and depth that match your role.
NumPy: numerical contracts before data analysis
Before starting: Python foundations, loops, functions and basic numerical reasoning. Computer typing and the native reader share material; actual NumPy execution uses the downloadable lab.
Build and explain: Construct aligned arrays, preserve ownership, choose explicit dtypes, account for missing populations and verify a reproducible bounded numerical report.
- 1. Arrays, dimensions and an explicit shape contract
- 2. Dtypes, promotion and widening before arithmetic
- 3. Indexing, slices and preserving a row axis
- 4. Boolean masks and compound selection
- 5. Copies, views and observable mutation
- 6. Reshape, transpose and meaningful axis order
- 7. Broadcasting and explicit column alignment
- 8. Reductions, axes and keepdims
- 9. Missing values, infinities and an honest denominator
- 10. Safe ufunc division with out and where
- 11. Stable sorting, tie rules and aligned identifiers
- 12. Unique labels, inverse indexes and grouped counts
- 13. Stacking features and computing a matrix product
- 14. Owned random generators and reproducible sampling
- 15. Numerical files and verified readback
- 16. Bounded numerical workspace and block processing
- 17. Reference-only scaling and held-out data
- 18. Capstone: an aligned technical-role report
Review evidence: Type short focused decisions, repair missing code and recall with the reference hidden. Then change one stated policy in a working local baseline and run array/dtype/shape, state and I/O acceptance checks. Saved learning progress is not professional certification.
Your learning path
0 of 18 lessons completed
Saved on this device
Module 1 · 1 lessons
Arrays, dimensions and an explicit shape contract
Construct a rectangular numerical array and distinguish shape, rank, size and Python container length.
Module 2 · 1 lessons
Dtypes, promotion and widening before arithmetic
Prevent fixed-width integer overflow by choosing the calculation dtype before performing the operation.
Module 3 · 1 lessons
Indexing, slices and preserving a row axis
Select rows and columns intentionally, including the difference between a scalar index and a one-column slice.
Module 4 · 1 lessons
Boolean masks and compound selection
Combine elementwise conditions with explicit parentheses and preserve the order of selected observations.
Module 5 · 1 lessons
Copies, views and observable mutation
Distinguish shared basic slices from independent advanced indexing and deliberate copies.
Module 6 · 1 lessons
Reshape, transpose and meaningful axis order
Change an array layout deliberately without confusing reshaping with transposition.
Module 7 · 1 lessons
Broadcasting and explicit column alignment
Apply one offset per feature without accidentally treating that vector as one value per observation.
Module 8 · 1 lessons
Reductions, axes and keepdims
Reduce the intended population while retaining the dimensions needed by the next operation.
Module 9 · 1 lessons
Missing values, infinities and an honest denominator
Report how many measurements were actually summarized and keep an empty result distinct from zero.
Module 10 · 1 lessons
Safe ufunc division with out and where
Avoid invalid arithmetic by controlling where the ufunc executes and initializing every output position.
Module 11 · 1 lessons
Stable sorting, tie rules and aligned identifiers
Sort scores through one shared ordering vector and make tie-breaking a deliberate policy.
Module 12 · 1 lessons
Unique labels, inverse indexes and grouped counts
Aggregate repeated labels without separating each measurement from its owning group.
Module 13 · 1 lessons
Stacking features and computing a matrix product
Construct aligned feature columns and distinguish a matrix-vector product from elementwise multiplication.
Module 14 · 1 lessons
Owned random generators and reproducible sampling
Sample unique identifiers without changing global random state or promising cross-version identical streams.
Module 15 · 1 lessons
Numerical files and verified readback
Persist and reload a numeric array in an owned temporary directory with an explicit non-pickle boundary.
Module 16 · 1 lessons
Bounded numerical workspace and block processing
Measure array bytes honestly and avoid constructing an unnecessary full-size intermediate.
Module 17 · 1 lessons
Reference-only scaling and held-out data
Estimate feature statistics from the reference matrix and reuse them without fitting to the evaluation population.
Module 18 · 1 lessons
Capstone: an aligned technical-role report
Combine canonical identity checks, one eligibility mask, stable ranking and population-consistent summaries.