I delivered a talk in PythonPH x Sykes about recommender systems. Personally, I find recommendation algorithms to be one of the most “human” of machine learning fields. You might like A, B, C because of D, E, F. Repeat for all users. Fun! And computationally expensive, I might add.
This talk is a rapid-fire survey of many basic recommendation algorithms. Here’s the slides.
Here’s the live video. You should check out the 4 talks which are all about data science topics.
Some of the featured libraries are the ever-trusty Numpy & Pandas. There’s also scikit-learn, surprise and Ben Frederickson’s excellent library, implicit.