Abstract:
Shopping centers play an important role in the traffic impact of any city. The proliferation
of shopping malls, particularly multi-purpose shopping centers (trans-marts) has caused a
significant change in traffic patterns in the cities. To study the travel patterns of these shopping
centers, there is a need for proper data and models for the travel demand forecasting. This will
ultimately help in the utilization of optimum resources with maximum benefits. The modern
revolutionized world requires planning, which necessitates the use of real data, models, travel
patterns, and trip generation models.
This study bridged the gap by developing the trip attraction rates of two metropolitan
cities. First, shopping centers with multiple facilities such as shopping, dining, restaurants, play
areas, and cinemas are chosen for research in Islamabad and Lahore. Six shopping centers are
selected, and data is collected every 15 minutes during the three peak hours on weekdays and
weekends. For the collection of data, two days on both weekends and weekdays are selected.
The data collected includes the number of people and vehicles entering and leaving the
shopping centers for every 15 minutes intervals. The data related to physical features like gross
floor area, shopping area, playing area, watching area, dining area, number of shops, number
of parking spaces and number of stories is collected from the management of shopping centers.
Different statistical techniques like multiple linear regression followed by non- linear
regression, partial least square regression, and artificial neural networks are used. All the
models gave significant results. Pearson Correlations showed that all the explanatory variables
have significant correlations with response variables except for the number of stories of
shopping centers. The data is observed to be nonlinear and multicollinear
