Machine Learning in Python or R?

MayCaesar

Python is arguably a much more universal language, and quite a bit more popular. With R, you are more or less stuck doing data science, while knowing Python allows for a much wider array of applications. Python also has absolutely unparalleled visualisation libraries.

Bear in mind that I do not have much experience with R, while Python is my primary language, so I may be biased. But as someone who has been doing data science for the last few years, let me assure you that Python is more than adequate for virtually any data science application. And machine learning in Python is very powerful, with dozens libraries available. I personally have delved in TensorFlow and Keras, and the latter is absolutely amazing.

BonitaVanhooser

MayCaesar said:

Python is arguably a much more universal language, and quite a bit more popular. With R, you are more or less stuck doing data science, while knowing Python allows for a much wider array of applications. Python also has absolutely unparalleled visualisation libraries.

Bear in mind that I do not have much experience with R, while Python is my primary language, so I may be biased. But as someone who has been doing data science for the last few years, let me assure you that Python is more than adequate for virtually any data science application. And machine learning in Python is very powerful, with dozens libraries available. I personally have delved in TensorFlow and Keras, and the latter is absolutely amazing.

I also preferring Python because it's more general if you talk about its use. However, after Python, maybe it could be an idea to try for R. Need one more suggestion, do you think online courses like free ones at coursea or udemy and other places are enough good? Or should I go with paid ones? I was willing to start with free online courses because I am just a beginner. What're your views?

MayCaesar

@BonitaVanhooser

It is hard for me to give suggestions as far as courses go, since I have never taken one myself: I switched from C++ to Python when I started my PhD program and learned the language on the fly. I have used mostly random online guides and source code to learn the language. In my opinion, you learn the language much better when you are working on something, than when you just study it in the abstract manner.

If you want to learn the language in depth, I would buy a solid comprehensive book on it, such as this one: https://www.amazon.com/Python-Data-Analysis-Wrangling-IPython-ebook/dp/B075X4LT6K And, of course, taking a solid online course would not hurt. I am not sure what benefits paid ones have over free ones; I would try out a couple of free ones and see if they work for you.

As far as real world problems to serve as exercises go, this is the primary website: https://www.kaggle.com/ Bear in mind that it is a very serious commitment, and most problems take a while to finish. But it is probably the most practical way to learn, as you will be solving the same problems as data scientists in real companies do. Practice on Kaggle can easily be converted into a strong resume, and employers highly value it.