Epidemiology, since its birth as a biomedical discipline to the present day, has progressively enriched and refined tools (and methodology) for research and generating evidence, adapting over time and in the context in which evidence is generated. In contemporary times, characterized by technological pervasiveness, increased computing power, and a global pandemic in an interconnected world, the paradigms of epidemiological research are opening up, at varying speeds depending on their real applicability, to a new, broader conception of "data" and more generally of its treatment. In this overview, we aim to take stock of what we know about this moment in epidemiology, where new research strands and "data-driven" analysis techniques are emerging alongside the traditional etiological vocation; a complex, ever-evolving scenario made up of lights, shadows, stimuli, and failures in which issues of method validity, professional training, and patients' right to privacy are increasingly central. The review therefore provides a starting point for reflection on this transition, identifying examples that support both the methodological and academic debate and case studies regarding the impact of big data in real clinical practice and, more generally, in service epidemiology.