What's a good way to go from almost 0 knowledge in Machine Learning to at least vaguely knowing what I'm doing? I tried the "Supervised Machine Learning: Regression and Classification" class on Coursera, and bounced off. I got some of the actual concepts, but whenever it got to the point of "now write some code", I wasn't able to manage, and found it difficult to gain knowledge from the lectures.

I have a background in software development, with most day-to-day development in Python. I'd like something that would take less than a couple weeks to get something useful from.