HackMD
How to use different platforms to learn Data Science?
Youtube
Content on youtube ranges from short "quick-trick" videos to 24-hour tutorial playlists.
For the former if there is some topic or tool you are unfamiliar with ("What is REDIS?") youtuber Fireship is great at giving 100 second overview on many tools and concepts:
- SQL in 100 seconds: https://www.youtube.com/watch?v=zsjvFFKOm3c
- REST API in 100 seconds: https://www.youtube.com/watch?v=-MTSQjw5DrM
- GraphQL in 100 seconds: https://www.youtube.com/watch?v=eIQh02xuVw4
Some of his videos assume some prior knowledge, but he still does great job making the concepts accessible in "100 seconds" format.
For something more focused and in-depth Rob Mulla's long-form Pandas content is also pretty good, he assumes viewers have basic knowledge of Python needed to follow along:
https://www.youtube.com/watch?v=DkjCaAMBGWM
For more advanced folks, he has neat series on quick tricks to make Pandas code more idiomatic:
https://www.youtube.com/watch?v=_gaAoJBMJ_Q&t=104s
Github
Reading code written by other people is probably fastest way to level-up and learn new tricks and useful patterns.
When trying something new I usually try to do it myself using tutorials/documentation but if I find myself hitting a roadblock I often find it productive to visit Github to see how others are trying to solve the problem I am struggling with.
Kaggle
Platform for Machine Learning contests + rich tutorials.
Main value for everyone are datasets and Notebooks sections that have a lot examples on how to do Exploratory Data Analysis and generally "wrangle some data" into useful format.
Free kaggle cloud-environments, called kernels are also decent alternative to google collab.
Books
Books quickly become outdated as software changes fast, but good books try to teach fundamental concepts that don't change as often.
Everyone can write a programming book, and everyone does so end results are of varied quality, but there are two reputable publishers:
If you pick random book from either, there is good chance it is actually good! Both publishers offer a lot of books for free, so search web before buying!
Note: O'Reilly is a publisher but also an online library that holds books from other publishers (like Packt) which aren't quite as well-written.
[TODO - some specific Python/Pandas books (?)]
