Dark matter makes up most of the mass of the Universe but remains mysterious. I discuss recent progress in constraining its properties by measuring its distribution in the Universe from tiny dwarf galaxies to giant galaxy clusters, and comparing this with numerical simulations. The latest results favour a cold, collisionless particle that must lie beyond the standard model of particle physics. I discuss the known small scale problems with this model: the cusp-core and missing satellites problems, and I argue that these are likely due to baryonic "feedback" during galaxy formation. I conclude with a discussion of experiments underway to detect dark matter particles, and the role that astrophysics has to play in these too. There is an exciting a very real prospect of detecting a dark matter particle in the next five years.