In this post, I want to show how I use NLTK for preprocessing and tokenization, but then apply machine learning techniques (e.g. building a linear SVM using stochastic gradient descent) using Scikit-Learn.