Screw Python

Title says all. I was forced to use Python for last year's machine learning project because that's what the Tensorflow/Keras API was written in. Now that I have access MATLAB/Octave (courtesy of Andrew Ng through his machine learning course on Coursera, thank you so much), I am no longer dependent on Tensorflow, and therefore, the Python programming language. What a relief. Even though it touts itself as a high-level, diversified tool for all things data science, I don't find its syntax or its implementation particularly intuitive or efficient. Perhaps that's just because I have a background in Java. 

But anyways, the reason I'm talking about this is because I've started on data normalization over the past week. I already have a nice blueprint for the neural networks I'll be using for phishing email detection, so all that's left is to train and test them. Unfortunately, this is turning out to be more difficult than expected. There are a lot of different ways I could parse the raw email data. So much information in the headers, so many different ways to extract meaning, and so many features to choose from. Here comes the hard part of the project.