Recent advances in data science are revolutionizing how ecologists explore and interpret complex ecosystem dynamics. This talk provides an overview of the evolution of data-driven approaches in ecology—from traditional statistical and mechanistic modeling to modern machine learning techniques and the emerging class of foundation models. Beyond analytical methods, we will emphasize the growing importance of global ecological datasets—including biodiversity monitoring networks, remote sensing archives, and open-access repositories—that enable large-scale synthesis and model training. Special focus will be placed on time-series foundation models, such as Chronos, which offer new possibilities for integrating heterogeneous ecological and environmental data to enhance forecasting and system-level understanding. By linking global data infrastructures with advanced modeling frameworks, this talk will illustrate how the convergence of data-centric and model-centric paradigms is transforming ecological prediction, understanding, and applications for ecosystem management.