This dissertation explores how we can better understand and influence systems, which are collections of interrelated or interacting elements that are organized in a way to achieve something. Systems are everywhere, from the global economy and natural ecosystems to mental health and the spread of infectious diseases, and they can be understood from different perspectives. The first part of this dissertation examines systems from a statistical perspective, presenting novel Bayesian tests that improve inference from empirical data. The second part examines systems from a causal perspective, outlining key concepts and tools for modeling the outcome of interventions. The third part examines systems from a dynamical perspective, highlighting the value of developing models to explore the temporal evolution of systems. This part also considers the potential for tipping points in systems, which are critical thresholds that, once crossed, lead to lasting and difficult-to-reverse changes. The practical value of generic early warning signals — derived from dynamical systems theory — for anticipating tipping points is assessed using simulation studies and empirical examples. The fourth and final part of this dissertation argues for an action-based perspective that aims not only to understand systems, but also to actively change them for the better. Using the climate crisis as an example, this part outlines ways in which both scientists and citizens can become more involved in securing a livable and sustainable future for all.