A free, self-contained textbook for learning Python, data science, machine learning, and AI, from first principles to production. Every chapter runs in Google Colab with no local setup required. The entire book is built around a single real-world dataset: the Stack Overflow 2025 Developer Survey.
Click the badge above to open the interactive Table of Contents, or jump directly to any chapter using the links below. All notebooks run in Google Colab, a free Google account is all you need.
No installation. No configuration. Open a notebook and run.
The book is structured in four parts across 13 chapters and 10 appendices.
Build a complete Python foundation from scratch. No prior experience assumed.
| Chapter | Title | Open |
|---|---|---|
| 0 | Orientation & Setup | |
| 1 | Python Fundamentals | |
| 2 | Intermediate Python | |
Master NumPy, Pandas, and the Python visualisation stack.
| Chapter | Title | Open |
|---|---|---|
| 3 | NumPy and Pandas | |
| 4 | Data Visualisation | |
Build and interpret ML and AI models across three data modalities: tabular, text, and images. Covers responsible AI and adversarial robustness.
GPU recommended for Chapters 7, 8, and 9. In Colab:
Runtime → Change Runtime Type → T4 GPU
| Chapter | Title | Open |
|---|---|---|
| 5 | SciPy and Statistical Computing | |
| 6 | Machine Learning with scikit-learn | |
| 7 | Deep Learning with PyTorch | |
| 8 | NLP and Transformers | |
| 9 | Computer Vision with PyTorch | |
| 10 | Ethics, Bias, and Responsible AI | |
| 11 | Adversarial ML and Model Security | |
Take a trained model from notebook to production.
| Chapter | Title | Open |
|---|---|---|
| 12 | MLOps and Production ML | |
Ten reference appendices covering local environment setup, Keras, SQL, Git, Docker, security, and more.
| Title | Open | |
|---|---|---|
| A | Python Environment Setup | |
| B | Keras 3 Companion | |
| C | Project Ideas and Further Reading | |
| D | Reinforcement Learning Foundations | |
| E | SQL for Data Scientists | |
| F | Git and GitHub for ML | |
| G | Docker and Containerisation | |
| H | MLSecOps: Securing the ML Pipeline | |
| I | Failure Modes and Troubleshooting | |
| J | Python Foundations: A Deep Dive | |
One dataset, start to finish. The Stack Overflow 2025 Developer Survey runs through every chapter as a continuous project thread, you're not switching to toy examples every few pages. By Chapter 12, you've built, audited, secured, and deployed the same salary prediction system you started in Chapter 6.
Depth where most books skip. Most ML textbooks end at model training. This one covers what comes after: fairness auditing (Chapter 10), adversarial robustness and red teaming (Chapter 11), MLOps and drift monitoring (Chapter 12), supply chain security (Appendix H), and a dedicated troubleshooting guide for the bugs that don't crash your code but ruin your models (Appendix I).
Built for active learning. Every chapter contains concept-check questions with collapsible answers (69 total), and three coding exercises at progressive difficulty levels — guided scaffolds, applied challenges, and open-ended extensions (39 total, all using the SO 2025 dataset).
Zero setup. Everything runs in Google Colab. Free T4 GPU access covers all deep learning chapters.
| Background | Where to start |
|---|---|
| No programming experience | Chapter 0 |
| Some Python, new to data science | Chapter 3 |
| Python + data science, new to ML | Chapter 6 |
| ML background, learning deep learning | Chapter 7 |
| Want NLP / transformers / RAG | Chapter 8 |
| Want computer vision | Chapter 9 |
| Preparing for ML engineering interviews | Chapters 6, 7, 11, 12 |
| Focused on AI safety and responsible AI | Chapters 10, 11, Appendix H |
- A Google account (for Google Colab)
- No local Python installation required
- Chapters 0–5: no prior programming assumed
- Chapters 6–12: basic Python familiarity helpful (Chapters 1–2 cover this)
All chapters use the Stack Overflow 2025 Developer Survey: a real-world dataset of ~15,000 curated responses covering programming languages, salaries, tools, education, and demographics. The dataset is included in the /data directory and loaded automatically in each notebook.
Complete beginner (53–68 hours): Chapters 0 → 12 in order, plus appendices as needed.
Python-literate, new to ML (30–40 hours): Start at Chapter 3, skim Chapters 1–2 as reference.
ML practitioner adding depth (15–20 hours): Chapters 7–12 + Appendices H and I.
Responsible AI focus: Chapters 6, 10, 11 + Appendix H.
Production / MLOps focus: Chapters 6, 7, 12 + Appendices G, H, I.
Code: MIT License
Content: Creative Commons Attribution 4.0 (CC BY 4.0)
You are free to use, share, and adapt this material for any purpose, including commercial use, with attribution.
Found a bug, a broken Colab link, or a better explanation? Issues and pull requests are welcome. Please open an issue before submitting a large pull request so we can discuss the change first.
Built with the Stack Overflow 2025 Developer Survey dataset.
All notebooks run in Google Colab. No local setup required.