In a galaxy, chemical enrichment takes place in an inhomogeneous fashion, and the Galactic Halo is one of the places where the inhomogeneous effects are imprinted and can be constrained from observations. I show this using my chemodynamical simulations of Milky Way type galaxies. The scatter in the elemental abundances originate from radial migration, merging/accretion of satellite galaxies, local variation of star formation and chemical enrichment, and intrinsic variation of nucleosynthesis yields. In the simulated galaxies, there is no strong age-metallicity relation. This means that the most metal-poor stars are not always the oldest stars, and can be formed in chemically unevolved clouds at later times. The long-lifetime sources of chemical enrichment such as asymptotic giant branch stars or neutron star mergers can contribute at low metallicities. The intrinsic variation of yields are important in the early Universe or metal-poor systems such as in the Galactic halo. The carbon enhancement of extremely metal-poor (EMP) stars can be best explained by faint supernovae, the low [α/Fe] ratios in some EMP stars naturally arise from low-mass (~ 13 - 15M ⊙) supernovae, and finally, the [α/Fe] knee in dwarf spheroidal galaxies can be produced by subclasses of Type Ia supernovae such as SN 2002cx-like objects and sub-Chandrasekhar mass explosions.