We introduce explicitly spatial models in the regression framework, including spatial fixed effects, spatial regimes, spatial lag, spatial error, and geographically-weighted regression, to predict and explain urban phenomena while accounting for spatial heterogeneity and diffusion.