Most works on the modeling of energy consumption in buildings make little headway in producing results that can be put to use in public policies. They reduce energy consumption to a quantity, without taking account of the behavioral processes and spatial variations that generate and explain them. The objective of our study is to reverse the perspective and to start from energy-related behavior rather than from intensities of consumption. Using regression methods and data from a survey on the energy lifestyles of 1950 households in the Île-de-France (Paris) region, we first model energy-related behaviors by looking at the characteristics of dwellings and buildings, as well as the socioeconomic and demographic characteristics of their occupants. Second, we in-vestigate the relationship between behavior and energy consumption by using our models to simulate domestic energy consumption at national scale. Finally, we map energy behaviors at the scale of the Île-de-France region. Energy-related behaviors turn out to be the product of complex interactions between habitat, inhabitants, and environment. Their spatial distribution is linked to territorial dynamics (metropolitanization, distribution of habitat types, etc.), as suggested by our cartographic analysis. But these results also highlight the complexity of the processes, insofar as there are no mechanical explanatory variables for behaviors.