Modeling the effects of land use characteristics on mode choice for home-based work trips: The case of Istanbul

Abstract

The cities in Turkey have been facing some of the same problems that European and North American cities have, including traffic congestion, traffic accidents, and air pollution. To overcome this situation, both local and central administrators who make urban policies and city planners have tended to optimize Land Use and Transportation Interaction (LU&T). In recent years, some new concepts concerning urban planning have suggested that shaping travel demands can be used as a tool to overcome these problems. The most common objectives of this concept are to reduce motorized trips and to promote public transit in metropolitan areas. To achieve these objectives, understanding the probable effects of land use on mode choice is crucial. However, the effects of land use on mode choice have never been answered fully, in Turkey. This dissertation empirically explores the relationship between travel mode choice and land use by employing different mode choice models for home - based work (HBW) trips in Istanbul at aggregate and disaggregate levels. The focus of this study is to understand how land use characteristics affect home - based work mode choice in the case of Istanbul. In this study, logit models and Bayesian Belief Networks (BBNs) are used to identify and quantify the effects of land use on travel mode choice at both levels. Empirical data were obtained from 2006 Household Travel Survey prepared for 2007 Istanbul Transportation Master Plan Study. The model results show that land use variables are statistically significant at both levels. The inclusion of land use variables increases models' explanatory level.