With the development of new detectors, telescopes and computational facilities, survey astrophysics is entering an era where deep optical surveys are possible for a large fraction of the visible sky. One of the largest of these surveys, the Large Synoptic Survey Telescope (LSST), will comprise an 8.4 m primary mirror with a 9.6 square degree field-of-view and a 3.2 Gigapixel camera. Over the ten years of its operation, the LSST will survey half of the sky in six optical colors, discovering 37 billion stars and galaxies and detecting about 10 million variable or transient sources every night. In this talk I will briefly describe the potential of the LSST for studying the nature of dark matter and dark energy, for measuring the properties of our Galaxy, and for creating a census of our Solar System. I will then focus on some of the challenges we will face when analyzing data that, while massive in size, is intrinsically noisy and incomplete and how we are using simulations and models to address the question of how can we optimize the science of the LSST.