Intro to Data Science & Machine Learning
This course is for students who want to build practical data science skills. You will learn to run a full data science project from question to conclusion, guided by the CRISP-DM process throughout. By the end, you'll have gathered experience in - discovering worthwhile challenges to address (innovation mindset), and - cleaning and exploring real datasets, as well as training and evaluating models, - and communicate results honestly and well with stakeholders.
The technical layer is only a middle layer. The before and after are at least as important.
Parts
| Part | Topics | Nuggets |
|---|---|---|
| Part I: The Big Picture | AI and data science definitions, CRISP-DM process map, academia vs. business DS | 5 |
| Part II: Personal Data-Science Projects | Why a personal project, idea generation, feasibility checking, scoping a result | 4 |
| Part III: Data Understanding | Data work reality, attribute types, exploratory data analysis (EDA) | 8 |
| Part IV: Data Preparation | Train-test splits, transformations, cleaning and encoding, data-processing pipelines | 6 |
| Part V: 1st Pass: Supervised Learning | Regression, decision trees, classification evaluation, random forests | 10 |
| Part VI: Principles That Transfer (Reflection) | Generalization, simplicity, baselines, aligning metrics, explainability | 5 |
TBD
Further Notes
About Learning: Becoming Good at Something
When learning something new and challenging, consistency is key.
There's no magic abbreviation to learning skills.
- Even "intelligent" AI tools don't provide a shortcut to your personal skill mastery, which needs to be earned the classic way: taking on the cognitive challenges that the learning journey brings to the table.
- This is true for this course and data science. In fact, it is true for any skill.
Grow beyond yourself. But while doing so, it's important to enjoy the journey.
How to Read This Course
Linear Path: Following the parts in order, nuggets in sequence within each part. This is the safest path through the material.
Reading with a Goal:
- For project reference, use the Part indexes to find the nugget that covers your current problem step.
- For exam preparation... well, we'll come to that.
AI Usage: Transparency Note (human-written)
Parts of this course have been drafted with AI support. When using AI tools, I aim at carefully reviewing content and adding my personal human notch.
How I work with AI and what I use it for: - I let it execute steps following precise instructions. - AI usually generates a first draft of prose, which I then review. This happens only after I fix structure. - I also used AI to program autoformatting tools. This includes automation for creating navigation bars, building table of contents as well as part indices, and checks for consistency (e.g., cross link audits, media, section audits): All of that is governed by Python scripts.
What is not AI: - The overall course structure. It's guided by purely human intent. I baked into this course what my experience called out to be necessary for practical learning.
Notify me if you discover errors: In any case, this script may contain errors. I take full responsibility. If you notice errors or inconsistencies, please let me know. I'm happy to revise.
Using AI to write, draft, and edit texts is a process that needs careful consideration and rigorous quality checks. It's much like what you'll learn yourself in this course: For anything to be done, the process matters a lot. A good process is a gate for quality. Done carelessly, it's easy to arrive at what's called "AI slop". I'm trying to teach to choose sides wisely. I hope you'll find this script to be a good example of human-AI collaboration.
AI tools will shape the future of work. Let's try to use them in a good way.
As always: Happy learning, happy life! 🫶