This study investigates consumer choice of shopping centre in the context of individuals' daily activity schedules. The hypothesis tested is that activity schedules may constrain location choice and moderate preferences for locations. Observations are derived from an activity-diary data-set and a constraint-based model is used to identify the choice set in each case. A choice is described in terms of a heuristic selecting a location from a choice set. The method CART is used to find the decision tree that best fits the choice observations using a large set of predictors. The results indicate that schedules have a significant impact on location choice primarily by limiting the available time-window in a substantial proportion of cases and secondarily by moderating preferences. It is concluded therefore that spatial shopping models used for analysis or prediction should be based on activity-constrained choice sets at an individual level.