Practical Data Science in Python: Guidebook

Caveat: It uses the worst possible technique for spam filtering: Naive Bayes, responsible for extremely poor spam filtering systems with tons of false positives and false negatives, still alive today. So this is definitely not a good resource to learn data science, but a great tutorial to learn Python, especially since naive Bayes is extremely easy to implement, though alternate but far better techniques such as hidden decision trees, are almost just as easy to code.