Two point galaxy clustering statistics are a key observation for large scale structure surveys. Unfortunately, it is very difficult to derive an accurate galaxy power spectrum from the halo model and perturbation theory. In this talk, I will be presenting a new framework for calculating the galaxy power spectrum, called emulation, which can make predictions accurate to ~3%. Our framework uses a set of fully non-linear galaxy power spectra measured from applying the halo occupation model on a dark matter halo catalogue identified in a LCDM N-body simulation. A Gaussian Process is then used to explore the space of halo occupation parameters. I will also describe a number of other projects involving the cosmic emulator framework, such as the redshift space power spectrum.