Modeling space preferences for accurate occupancy prediction during the design phase

The accurate prediction of occupancy during the design phase of a building helps architects to improve space efficiency by eliminating the possible under-utilization and over-crowding of space during the design use phase. However, existing models exhibit limited accuracy in occupancy prediction. A major reason for this limitation is that spatial-choice behavior is ignored or oversimplified. We therefore developed a space-preference model to explain spatial-choice behavior, with a particular focus on individual work-related activities. For this purpose, we conducted a discrete-choice experiment: 2048 observations of spatial choices were collected, and a conditional logit model was used to model space preferences. The application of the space-preference model was illustrated by two case examples, with which the merits of the model were highlighted. It was then validated in a predictive success test and a case study. The model will help architects to assess potential over-crowding and under-utilization of space according to different design options.