Data Science
Lab Curriculum.

BUG FOUND: graduates who can run a notebook, but cannot defend an analysis.

$ step-01 brushup // weeks 1–2 — Topics taught:

1. Descriptive statistics: mean, median, spread, and why each can mislead.
2. Python for analysis: pandas, plotting, and reproducible notebooks.
3. Exploratory analysis: distributions, outliers, correlations, and missing data.
4. Hypothesis testing basics and common misreadings (p-values, significance).
5. Regression basics and residual checks.
6. Writing an analysis so someone else can reproduce it: data, code, and assumptions.

$ step-02 debug // weeks 3–4 — Production Debugging. Tools: Python, Jupyter, SQL, plotting libraries.

• BUG #1: A chart claims a trend that disappears after cleaning — find the outlier / aggregation error, correct it, show both versions.
• BUG #2: A test is applied to the wrong data type / design — choose the correct test, rerun, explain the change.
• BUG #3: A biased sample drives a wrong conclusion — identify the sampling bias, reweight or resample correctly, state the limit.
• BUG #4: Missing data handled by silently dropping rows — compare methods, choose a defensible one, show the impact.

$ step-03 navigate // weeks 5–7 — Codebase / System Navigation:

• Reproduce someone else's notebook end-to-end from raw data; fix broken paths, versions, and undocumented steps.
• Correct an analysis that mixes correlation with causation; rewrite the conclusion with evidence limits.
• Audit a dataset for duplicates, label errors, and leakage before modelling.
• Present an analysis to a mentor as if to a non-technical stakeholder: one claim, one chart, one caveat.

$ step-04 ship // weeks 8–9 — Shipping Features: capstone — a validated analysis + model report answering a real question, with methods, limitations, and next actions, presented live.

$ ship → PROOF — exact output: reproducible notebook, analysis report with methods and limitations, and a one-page executive summary.

$ run --lab data-science → ✓ SHIPPED, job-ready