The depth of the course: 15 notebooks, roughly 12–14 hours of focused work. It revisits everything from the 90-minute Introduction Session in substantially more detail, then goes far beyond it — data cleaning, visualisation with matplotlib and seaborn, exploratory data analysis, machine learning with scikit-learn, and an introduction to PyTorch, finishing with an end-to-end capstone project.
Designed for independent study. Each notebook builds on the previous ones and states what it assumes, so you are never starting from zero — but there is enough explanation and worked example to learn on your own, without an instructor.
☁️ Colab tip: notebooks open read-only from GitHub — click File → Save a copy in Drive once at the start so your exercise solutions persist. Runtime → Run all works in every notebook. Each notebook also carries its own Open in Colab badge in the first cell.
Revisits Track 1's foundation and adds the depth a practitioner needs.
| # | Notebook | Time | Covers |
|---|---|---|---|
| 1 |  Python Basics |
30–35 min | Types and conversion pitfalls, rounding surprises, float precision, strings |
| 2 | Control Structures |
35–40 min | if/elif/else, loops, convergence loops, break/continue, try/except |
| 3 | Lists and Sequences |
30–40 min | Indexing, the slicing model, comprehensions, zip, tuples, aliasing |
| 4 | Dictionaries and Nested Data |
35–45 min | Key-value lookup, nested data, list-of-dicts as a table, counting, JSON |
| 5 | Functions and Modules |
35–40 min | Parameters, defaults, return vs print, *args/**kwargs, scope, imports |
| # | Notebook | Time | Covers |
|---|---|---|---|
| 6 | NumPy Fundamentals |
45–55 min | Arrays, vectorisation, broadcasting, axes, reproducible randomness |
| 7 | Pandas Essentials |
45–55 min | Series, DataFrames, loc/iloc, boolean masks, groupby |
| 8 | Data Cleaning and Preprocessing |
45–55 min | Missing values, dtypes, duplicates, categories, outliers, scaling, encoding |
| 9 | Visualisation with Matplotlib |
50–60 min | Figure/Axes model, choosing the right chart, subplots, annotations |
| 10 | Visualisation with Seaborn |
50–60 min | Long data, axes- vs figure-level, distributions, categorical plots, faceting, heatmaps, palettes |
| 11 | Exploratory Data Analysis |
50–60 min | The EDA workflow on a real dataset: distributions, relationships, correlation |
| # | Notebook | Time | Covers |
|---|---|---|---|
| 12 | Machine Learning Basics |
45–55 min | What ML is, supervised vs unsupervised, features/target, train/test, evaluation, overfitting |
| 13 | The Scikit-Learn Workflow |
70–85 min | Classification and regression in practice, pipelines, metrics, GridSearchCV |
| 14 | PyTorch Basics |
50–60 min | Tensors, autograd, a small neural network, the training loop |
| # | Notebook | Time | Covers |
|---|---|---|---|
| 15 | Capstone Project |
75–105 min | End-to-end project: data, EDA, dashboard, regression, executive summary |
Work through the notebooks in order — each assumes the previous ones.
| Pace | Plan |
|---|---|
| 1 notebook / day | Done in about three weeks |
| 3–4 notebooks / weekend | Done in about four weekends |
| Intensive week | The whole track in 4–5 focused days |
If you attended the live session, notebooks 1–7, 9 and 10 will feel partly familiar by design: each opens with a short "🔗 Building on Track 1" note telling you what is recap and what is new, so you can move quickly through the parts you know.
- Header — track position, time estimate, learning objectives, prerequisites.
- A "🔗 Building on Track 1" note where the topic was introduced in the live session.
- Intuition first, then code, then interpretation of what the output means.
- Exercises with complete solutions (collapsed), including a "Debug me 🐞" challenge.
- Key takeaways, a self-assessment checklist, and a pointer to the next notebook.
Continue with the summer school's Hands-On-Notebooks — a visual, intuition-driven collection on linear models, decision trees, random forests, and gradient boosting.