Abstract: I will introduce several methods to extract cosmological information from observations of the sky, from the cosmic microwave background (CMB) to hydrogen line intensity mapping. Starting with the first light, the CMB, a cosmological Boltzmann code is presented to calculate the CMB power spectra from different theoretical models, for a flat $(\Omega_k = 0)$ cosmology. It can rapidly compute the CMB power spectrum accurately up to high multipoles. After reviewing the chronological evolution of the Universe, I will then introduce the basic ideas of intensity mapping of neutral hydrogen after recombination, which tells us about structure formation. I will present the first results from the Tianlai Dish interferometer array, which is a new instrument specifically designed and constructed for hydrogen intensity mapping between redshift $z=0$ and $z=2.55$. The array is still in its infancy, and a thorough understanding of the instrument through simulation, calibration, noise analysis is described. An eigen-decomposition method to remove the Sun signal from the timestream data of radio interferometers is discussed in detail. It is applied to Tianlai data, which helps the instrument collect usable data during the daytime. Finally, a machine learning method is presented that maximizes the cross-correlation between hydrogen intensity maps and galaxy redshift surveys as a tool for detecting the hydrogen signal in the presence of bright foregrounds.