Where to start your data science learning journey?
Anyone interested in learning data science should at least once watch Andrew Ng’s Machine Learning course on Coursera. Amongst all the training material available on the web, this is the one I find by far the most remarkable to get started in this field.
Andrew has an amazing ability to demystify all topics and turn them into something that seems simple even after watching the course videos for the first time, but would be totally confusing otherwise. And this is true this for audience with little or no prior knowledge of the area. If you consider the complexity of some of the topics covered, that’s quite a achievement.
What makes Andrew Ng’s course so remarkable?
What I admire in his teaching style and I believe make it so successul are the following elements:
- Great structure. There is a very logical flow in the sequence topics are being taught and concepts being introduced, and it helps with the learning process
- Great graphs. The frequent use of simple, well drawn graphs definitely helps a lot. In particular I find the use of color is pushed to the perfection. Don’t expect anything like Van Gogh though, the palette is a little bit more limited, with maybe just 3 or 4 different colors used. But even only 4 colors with such mastery really helps connecting the dots and understand which elements or phenomena are linked and which ones are not.
- Great explanations. What is said is just what you need to understand. Nothing to add, nothing to throw away, every word is at the right place. And slides are being annotated as the lecture progresses, so you don’t start with very verbose slides that you don’t understand, instead you get fewer information at the beginning and more is being added along, so it simplifies the learning process.
- It’s free. So if after watching the videos you realise data science is not your thing, you won’t have made too much of an investment to reach that conclusion.
And here comes the intuition
But what I like first and foremost, and to me what makes a real difference is that Andrew always starts with what he calls the intuition. In other words, for every lecture, he first focuses on what is the idea behind what will be explained. Why do we even ask ourselves the question we are trying to elucidate? Why is it useful? What is a simple real world example where we would meet this problem? Never, ever the course starts with details. The context is set first, and then he proceeds with details which are then much easier to apprehend.
This makes a huge difference. If every teacher in the world was to get into the habit of explaining the intuition before starting a course, regardless of the subject they teach, this world would definitely be a better place to learn!
So in essence, Andrew Ng teaches you machine learning, but if you are interested in teaching, you definitely have a lot to learn from him as well through these videos!
More great material
As a side note, another very good set of videos I recommend watching is the ones recorded by Hastie and Tibshirani based on their book “An Introduction to Statistical Learning”. They are also very clear, cover slightly different topics, and they have quite a bit of humor which I particularly enjoy (yes science can be fun).