Binary stars are common, and exert an outsize impact on many areas of astrophysics. They are the source of a wide range of phenomena from type Ia supernovae to gravitational wave sources. Despite their importance, the formation mechanism for the progenitors of interacting binaries are still poorly understood. In this talk I will describe our understanding of binary formation in general, and our recent progress on the origin of close binaries. I will emphasize the new statistical leverage provided by large surveys such as Gaia and SDSS IV. I will show how the evolution of the observed binary occurrence rates as a function of stellar mass, period, and metallicity can place tight constraints on formation models, and shed light on our understanding of exoplanet statistics as well.