Practice pre-processing different types of categorical features: https://www.kaggle.com/c/cat-in-the-dat/overview
I leaned on Pandas more heavily in this exercise than in the Titanic, all pre-processing of categorical fields was done within Pandas, and seems a little more straight-forward than doing the same in scikit-learn.
I also calculated precision, recall, and F1 scores for each variation of hyper-parameters, rather than just the accuracy score I'd used for Titanic.