Integrated Modeling of Urban Shopping Activities

Abstract

This paper presents an integrated modeling framework to examine the factors affecting urban out-of-home shopping activities. The framework consists of four integrated models: (1) children belief (binary logit) model; (2) mode choice (multiple logit) model; (3) duration (hazard) model; and (4) frequency (ordered probability) model. The output of the belief model, which captures the influence of children on shopping activities, is used as input to the other three models. The models were estimated using disaggregate data collected in Amman, Jordan, through personal interviews of 570 commuters. Frequency models were estimated for three categories of shopping activities: (1) during home-to-work trips; (2) during work-to-home trips; and (3) after work-to-home trips. The results demonstrated that the belief probability was primarily determined by household socioeconomic characteristics. Also, these characteristics and the belief probability were significant in the mode choice, duration, and frequency models. The mode choice was significantly influenced by the origin of shopping activity, type of postshopping activity, and travel time to shopping activity. The duration was significantly influenced by the type of postshopping activity, origin of shopping activity, and time of day. The frequency was significantly influenced by the travel distance from home to work. The results also showed that the frequency of shopping during different times of the day should be modeled separately, and that shopping duration and frequency are influenced by the commuter's belief in the children's influence.