Geospatial data available to researchers has increased tremendously over the last several decades, opening up opportunities to define residential location in multiple ways. is has led to a myriad of variables to define "location" in residential location choice models. In this paper, we propose a common classification for location variables and categorize findings from a wide range of studies. We find similar preferences but different measurement methods and market segments for locations across different study regions. Recent studies consider the residential unit as choice alternative, making it possible to include a detailed description of the built environment. However, these studies are still limited in number and the inclusion of socioeconomic environment is more common. Transport land-use models can benefit from the inclusion of points of interest, such as schools, network distances, and the distance to previous locations. For the results of location choice models to be transferable to different disciplines, and avoid multi-collinearity, it is necessary to present different model specifications, including variables of interest in different disciplines.