Billion-star surveys have transformed the Milky Way into a uniquely detailed laboratory for studying how disk galaxies form and evolve. From our position inside the Galactic disk, we can observe vast numbers of individual stars, each carrying information about the Galaxy’s chemical, dynamical, and temporal history. Ideally, reconstructing this history would require the ages, chemical compositions, and orbits of stars across the entire Galaxy. In practice, these quantities must be estimated from heterogeneous surveys that provide complementary but incomplete views of different stellar populations across Galactic regions. In this talk, I will present statistical and computational methods for integrating multi-survey, multi-modal, and multi-epoch datasets into a coherent view of the Milky Way. These methods include time-series analysis, physics-informed machine learning, and simulation-based inference, which connect observations to complex models of galaxy evolution. In particular, my talk is organised around three key questions: 1. How can we combine multi-survey observations to infer the physical properties of stars? 2. How can those properties be used to reconstruct the formation history of the Milky Way? 3. What insights can the Milky Way offer about galaxy evolution in general? I will show how combining spectroscopic, asteroseismic, and astrometric data yields precise estimates of stellar ages, chemical compositions, and orbits. I will then use these estimates to study secular (internal) processes in the Galactic disk and their connection to the Galactic bar. Finally, I will discuss how upcoming datasets, including Euclid and Rubin LSST, will extend this framework beyond the disk to the Galactic halo, enabling a more complete reconstruction of the Milky Way’s formation history.