1. The role of this book¶
This Jupyter Book is the central reference for SDS320. You’ll use it to follow technical examples, connect topics to your own project, and return to important instructions throughout the semester.
It isn’t meant to be read once, start to finish, and then forgotten — it’s closer to a project companion than a textbook. Some pages introduce concepts, some guide project decisions, some are technical reminders, and some help you check whether your work is ready for submission.
The book also contains more material than you’re expected to read cover to cover. That’s intentional. A good guiding principle:
Survey broadly, specialise narrowly.
Develop a general understanding of the methods introduced in SDS320, but study in depth the methods that are useful for your own project.
For example, you should understand roughly what object detection, semantic segmentation, change detection and regression can do — you don’t need to work through every technical page for every method. Let your own project decide where you go deeper.
2. Main parts of the book¶
The book is organised into several main sections:
Course – orientation material: course idea, schedule, assessment, AI use, resources and feedback.
Python & project setup – prepares your working environment and reactivates Python, notebooks, scripts and Git.
Project handbook – supports your independent project from first idea to final report and presentation.
Lessons – weekly topics and method-oriented workflows.
Project gallery – selected student projects from previous years, for inspiration.
Reference – glossary entries, cheatsheets, model-evaluation guidance, common errors, datasets and further reading.
3. How lessons are structured¶
Most lessons follow the same modular structure: an overview page introduces the topic and lays out a learning path, then focused chapter pages take you deeper.
A typical lesson includes:
an overview page with context, motivation, learning objectives and roadmap,
conceptual chapter pages,
workflow-oriented pages,
practical advice and common pitfalls,
a project-transfer page.
The project-transfer page is especially important — it’s what connects the lesson to your own SDS320 project.
Not every page needs the same level of attention. As a rough guide:
| Reading mode | What to do | Typical time |
|---|---|---|
| Core reading | Material that supports the current stage of almost every project. | 15–40 min |
| Method scan | Enough to understand what a method does, when it’s useful, and whether it could help your project. | 10–20 min |
| Method deep dive | The concepts, workflow, evaluation and limitations of a method you plan to use. | 1.5–3 h |
| Reference / revisit | Return to a page when you need to make a decision, solve a problem, or check your progress. | 5–20 min |
These times cover active reading and inspecting examples — not the coding, data wrangling or debugging that comes after.
4. Your weekly routine¶
A simple routine that works well:
Before class — spend about 15–30 minutes on the core reading for that week. Start with the lesson’s overview page: its context, motivation and learning objectives. If the topic looks relevant to your project, keep going through the pages that follow; if not, a general impression is enough.
During or after class — decide whether the week’s method is actually relevant to your project.
During project work — invest further time only in the methods and handbook pages that support your next project decision.
Useful preparation for a topic that interests you usually includes:
reading the relevant lesson pages,
running or inspecting the provided code examples,
noting down anything unclear,
thinking about how the topic connects to your project,
updating your project notes or repository if a lesson changes your plan.
You don’t need to understand every detail perfectly — a general idea, and maybe a few specific questions, is a good outcome for a first pass.
5. Using the book during project work¶
During independent project work you’ll often move away from the weekly topics. You might return to the Project handbook when refining your research question, planning your workflow, improving your repository or preparing your report. Or you might dip into Reference for a quick reminder on terminology, common errors, model evaluation or datasets — think of it as a workbench of tools rather than a set of chapters.
This is expected. Project work is iterative: it’s normal to revisit a decision you made weeks earlier, especially around your concept presentation and again before final submission.
6. When you feel lost¶
If you’re unsure where to start, go back to the parent page of the current lesson, check its roadmap, and find the chapter that addresses your question.
If the question is project-related, check the Project handbook first. If it’s technical, check the setup and troubleshooting pages. If you still need help, ask on MS Teams in the appropriate channel so others can benefit from the discussion too.
You can also use the feedback page to report unclear explanations, broken links or code that doesn’t work as expected.
7. Key takeaways¶
The book is a project companion, not something to read cover to cover.
Read the core material throughout the semester; scan the other methods; study deeply only what your project actually needs.
Return to the Project handbook and Reference whenever you need to make a decision or look something up.