Poisson regression is an effective machine learning technique that models counts or frequency data. Examples of this type of data include the number of insurance claims per year or rainfall events per month. We learn about generalized linear models in scikit-learn and how they connect to the PoissonRegressor. For tree-based models, we cover the algorithm behind scikit-learn's HistGradientBoostingRegressor, concentrating on Poisson loss. This talk reviews just enough theory to get an intuition for the various machine learning models and focuses on applying scikit-learn's Poisson regression techniques on a bike-share dataset.