During the early stages of the COVID pandemic there was a lot of discussion about flattening the curve, to prevent overload on our health care system. So what exactly is that curve, and how do you fit it to the raw case report data coming in? We will discuss several approaches and show you when and how they work. Once you understand the nature of outbreaks, you can start comparing them using the available context. As an example, we will try to make sense of the influence of measures such as lockdown and economic intervention using state of the art modeling and explanation tools such as XGBoost and SHAP. We will share all code and analysis with you as Python notebooks from our git repository so you can pick it up for further exploration. This session would be suitable for both interested and experienced participants, as an example of refining and modeling data to generate useful results. We will encounter many of the usual pitfalls and be able to discuss the typical solutions, providing an interesting insight into the daily practice of data science.