Examining children's weekday activity travel patterns and mode choice for school trips in Dhaka city, Bangladesh

Abstract

Analysis of travel pattern is an important research topic in the field of transportation engineering and urban planning, irrespective of developed and developing countries. It provides the background information necessary to better understand the complex relationship among urban structure, transportation system and people’s activity participation. The growing volume and complexity of urban travel in developing countries has become a major concern to transportation planners, service sponsors in urban areas, and policy makers. Designing transport strategies which meet the common political aims for the environment and the society requires a deeper insight into the routines of individual travel behavior. As a large proportion of population, children’s activity travel behavior is a paramount concern to be studied. On this regard, this study uses an activity-based travel diary survey dataset, which has been done between February 08, 2012 and February 28, 2012 and it generated a sample of 245 students in grades 6-10 in Dhaka Metropolitan Area, Bangladesh. In order to explore tradeoffs among factors that affect the frequency of out-of-home physical activity including the use of active transportation this paper utilizes a Zero-inflated Poisson (ZIP) regression model. Additionally, parametric hazard models are estimated to examine the duration of physical activities. The results of the ZIP count model suggest that personal and household characteristics, built environment attributes, and time commitments for other mandatory and discretionary activities influence the frequency of participation. For example, an increase in the number of children in households, and the presence of sidewalks, increase the frequency of out-of home out-of-school physical activities if everything else held equal. In contrast, travel time to school, presence of traffic intersections, and time commitment for tutoring sessions negatively affect activity participation. In terms of modeling duration of out-of-home physical activities, this study finds that a Weibull parametric hazard model outperforms a log-logistic model. The duration is influenced by socio-demographic characteristics, spatial context, and escort arrangements. For instance, car ownership and parental escorting increase the duration of physical activity. It offers an in-depth behavioral understanding of children’s physical activities, particularly in the context of a developing country, which is very limited in the existing literature. This study also examined mode choice behaviour for school trips in Dhaka City, one of the densest cities of the world. A conventional random utility-based multinomial logit model (MNL), a Mixed Logit (RPL) and a Latent Class Model (LCM) are used to examine the factors affecting mode choices. The choice set includes four different modes: car, three wheel Pedal-powered vehicles (TWPV), walk, and bus. These models were generated for school trips to determine factors affecting trip generation. The models results reveal that increase in travel time decreases the probability of choosing alternative modes. In addition, socio-economic factors such as household monthly income, home ownership, and number of children at household play vital role in choosing auto for school trips. Similarly, presence of intersection nearby the home neighbourhood reduces the probability of walking. It is also found that personal and household attributes, and transportation system characteristics has significant influence on children`s mode choices for school trips.