The formation of molecular gas is a critical step in the conversion of interstellar gas into stars, yet the physical processes involved still remain unclear. With a goal of providing observational constraints on the formation of molecular gas, I perform two high-resolution, multi-wavelength studies of the Perseus molecular cloud in the Milky Way. In the first study, I investigate the transition from atomic (HI) to molecular hydrogen (H2) on sub-parsec scales and find that the HI distribution is surprisingly uniform. As a result, the H2-to-HI ratio linearly increases with the total gas column density. These results are consistent with the theoretical model by Krumholz et al. (2009), where the formation and photodissociation of H2 are in balance and the abundance of H2 is controlled by the minimum HI column density required for H2 shielding. In the second study, I examine the relation between the H2 column density and the carbon monoxide (CO) integrated intensity and show that the ratio of the two, so called "X-factor", varies spatially by up to a factor of 100. I then compare the HI, H2, CO, and X-factor data with two contrasting theoretical models, i.e., PDR model by Wolfire et al. (2010) and MHD model by Shetty et al. (2011). I find that the steady state and equilibrium chemistry model (PDR) reproduces the observations very well but requires an extended halo around a dense core. While agreeing with the observations reasonably well, the macroturbulent and non-equilibrium chemistry model (MHD) shows interesting discrepancies, including a broader range of HI column density