Do you know Guido van Rossum? The Zen of Python? Do you know that the language name comes from the Monty Python comedy group? For a first introduction, I invite you to read the first two parts History and Features and philosophy from this Wikipedia link.
Here’s the initial learning pattern I followed:
- Python coding foundations
- Statistics and probability
- Introduction to data science
- Data visualization and data analysis
- Machine Learning
- Deep Learning
If your time is limited, it’s possible to skip the sections Statistics and probability and Deep learning. The idea here is that we can learn to drive without having to fully understand the mechanics of the car. This level of knowledge comes naturally with the time. For data science, it comes through reading (books, newsletters, or articles), by the questions you gonna ask yourself, and PRACTICE.
About the learning platforms:
- DATACAMP: Marketing is good and I’m sure the team is competent. But you have to pay for a year. And even if the app is nice, we get bored quickly. And the practical classes are not up to par. Write and re-write lines of code that don’t help to the understanding of coding, a bit mind-dumbing. I would not recommand. 2/5.
- EDX: There are quality classes. But the platform seems old, obsolete. Otherwise good. 3/5.
- UDEMY: The courses are sold almost 200 €. But regularly you have offers, and the courses worth ONLY 11 €. Be careful. I really loved this platform! 4.5/5.
- COURSERA: I loved. For many courses you need to pay a subscription after the free 7 day trial. In my case, I did the maximum during these free days and then unsubscribe. For learning the basics of Python, it’s okay to do it in that way. For more specialized courses, subscribing is the best option. The quality of the content is excellent on Coursera. 4.5/5.
- UDACITY: I did the free course Intro to data analysis, I loved it! The course and the platform. Probably my favorite! 5/5.
- COGNITIVECLASS: I did the DataScience Fundamentals course by IBM, it’s an excellent platform, and introduction to data science. 5/5.
Data science courses list
I propose this list of 10 courses with my personal opinion. This list allowed me to be autonomous and to code my first kernel on Kaggle, with the famous Titanic dataset. By taking notes, and practicing daily, give yourself 3 months to be ready for your first machine learning in Python.
The Deep Learning section isn’t in this list of 10 courses. If you’re interested in, you can find additional courses resources at the end of this article : Deep Learning, Tableau (software), Machine Learning, and Free Books
1. Python coding foundations
Review: The course is perfect to start. The teachers are great. We are learning the basics gradually. The course also provides summary sheets.
Review: This is the continuation of the preceding course. We discover the creation of functions, and good coding practices. Excellent course. Sometimes the tests are difficult to validate because of the multiple-response questions. And as the number of tests per day is limited, and we don’t have access to our answers on the old tests, my advice is to film your tests. It avoids losing too much time on the questionnaires, the true goal is to learn to code in Python.
Review: When you’re comfortable with the loops, and you understand the Python lists, it’s time to create your first programs. For this part, there is a 7 free day-trial, I recommend organizing your schedule to make a maximum. The code is fully accessible, if you can not finish do not hesitate to download the code to understand it later. You will code small games like this one « Guess The Number« , « Rolland Garros » (How to play: Player 1 (W / S) | Player2 (arrow down / arrow up) + Press play to launch, on up left).
Review: This course can be deleted because it’s more for fun than data science. But if you have time and want to code games like « Black Jack« , « Spaceship » or « Rice Rocks » it’s the perfect course!
2. Statistics and probability
Review: It’s a dense course, where you need to have some knowledge in probability or you’ll be quickly exceeded. The exercises are very good, excellent quality of the courses. Negative point, exam sessions are passed, and you can’t validate the responses of the exercises. It’s a shame but not embarrassing. About your learning program for data science, and the goal of coding your first machine learning, it’s okay to skip this course.
3. Introduction to data science
Review: Beautiful introduction. We can discover our first big picture of machine learning processes. The videos are done very well. Keep in mind that the mentioned tools can be outdated. The course is old, and the world of data science is constantly evolving.
Review: This is another introductory course in data science. The course is of good quality, good advice, and resources to go further. Attention, the course must be bought during an offer (9.99 € instead of 194.99 €).
4. Data visualization and data analysis
• Intro to Data Analysis (20h)
Review: You had the time to digest the code learned, improve your knowledge of statistics, and took your first steps in the world of data science. It was my favorite moment, start the data analysis in Python, with numpy and pandas libraries. I really fell in love with this course, and the Udacity platform. Excellent quality, with a well thought practice. Thanks to them for this free resource! Hope you’ll enjoy this part too.
Review: Another great course, with lots of advice. Here, you go further with Pandas (GroupBy, Datetime, and more). Boris Paskhaver is perfect. The course well built with quality resources. Congratulations! Attention, the course must be bought during an offer (9.99 € instead of 199.99 €).
5. Machine Learning
Review: The course is a little outdated, but it is still of great quality! Many codes are provided, slides, and explanatory videos. A good base to start in machine learning. I focused on Python and skip all R videos. Attention, the course must be bought during an offer (9.99 € instead of 199.99 €).
Here is the complementary list. However, I think that from a certain level, practice, Kaggle kernels and readings are the best way to learn. Also comes the time to specialize: Bokeh, Dash, NLP, Keras, GAN, knowledge of hyperparameters algorithms, … and at this time, there are tailor-made courses that you can find on the main learning platforms as the Tableau courses you can find below.
Machine learning course from Andrew Ng is famous, because of Andrew Ng and its quality. Don’t hesitate to subscribe on this Coursera course.
You can find e-books PDF, most about statistics. I recommend you the pleasant reading 25 Data Scientist Advices. You’ll find 25 interviews with the best worldwide data scientists, giving feedbacks. It’s interesting to read for both beginners and project managers.
- 25 Data Scientist Advices (recommended)
- Top 10 Data Scientist Mistakes
- The Elements of Statistical Learning
- Probability Theory, The Logic of Science
- Statistics – New Foundations, Toolbox, and ML Recipes
- Classification and Regression
- Foundation of Data Science
- Introduction to Statistical Learning (R)
If you have any resources you want to share, feel welcome !
- - -