Although many frameworks and libraries exist for machine learning research and for specific modeling approaches (e.g., reinforcement learning, geometric learning), more and more libraries are emerging that focus on common business tasks or data modalities – like clickstream analysis or log analytics. In this vein, Kats is a framework recently released by Facebook that helps practitioners perform time series analysis, including forecasting; detection (e.g. of seasonality outliers, slow trend changes); feature extraction and embedding; and multivariate analysis. Kats provides a one-stop shop where users can access both classical and advanced algorithms and models across these four domains.