Is it necessary to learn Python from square one if your ultimate goal is to use it for data science? I have used R a lot for data science. But could I learn Python sufficiently well just by looking at other data science problems and their Pythonic approaches?
Both Python and R are extensively used in the field of data science.Though both have their advantages and disadvantages, python has many libraries that are awesome to use for data science purpose. For example, scikit-learn, Keras, Theano, XGBoost, Matplotlib,etc. These make lives so easy for data science.
Yes, analyzing and understanding the code of others can give you a lit of information but you cannot learn until you don't get your hands dirty with some python code. In order to learn how to use python in data science, this link provides an excellent tutorial.
Python is a very intuitive language in terms of syntax. The language is designed keeping readability in mind thus is a very powerful and efficient language with a flat learning curve. R is a language with much stronger data analysis toolkit and support and a much steeper learning curve.
The pandas package in Python implements the R data frame and should be a good place to start.