In this post, I will collect some online data science resources such as online classes, blog posts etc. that I personally find very useful. (Ratings: ⭐ useful, ⭐⭐ very well done, ⭐⭐⭐ recommend absolutely)
Coursera: IBM Data Science Professional Certificate ⭐⭐
This class is a perfect introduction for anyone entering the field. It does not assume much prior knowledge, even though some prior programming experience is helpful. Even if most of it is not completely new material, as it was for me and probably for most people, it is a great refresher: If you already know some topics, you will be quicker going through exercises and labs, e.g. you won’t lose money because of it.
Pricing: around 35€ / month.
Pros:
- Very comprehensive: Covering theoretic methodology, SQL, Python tutorials, visualisation, etc.
- Nicely done quizzes and labs.
- The community is quite active in the forums: Your questions will be answered quickly.
- The class gives an overview into the IBM data science framework around Watson studio.
- Not very expensive. The more time you have, the less you spend.
Cons:
- Within the class, some of the individual classes are a bit redundant.
- Lecturers have varying presentational skills.
- The final project for the last class (Predicting car accident severity based on a police accident report data set) was less well curated.
- It does not go beyond IBM tools.
- There is little information or material on how to “production-ize” a machine learning project.
Coursera: Deep Learning Specialization (deeplearning.ai) ⭐⭐⭐
This class picks up where the IBM classes stopped. Once you have mastered the foundations in machine learning such as logistic regression, you start with the more complex concepts such as neural networks. I find this class extremely well curated. This doesn’t as a surprise, since Andrew Ng, the course lecturer, is also the co-founder of coursera, so it makes sense that he is able to leverage this platform to its fullest.
Pricing: 41€/month
Pros:
- Andrew Ng’s tutorial videos are great extremely well done and nice to follow.
- The labs are interactive (auto correcting) and figure into your final grade. (Usually in coursera, it’s quizzes and peer-reviewed assignments)
- The community support by the organizer team is very quick and comprehensive.
Cons:
- The forums are a bit spammed sometimes and it is sometimes hard to find a relevant post. This is probably a general Coursera problem.
I’m currently enrolled in this class, so this verdict is not final.