Influence of Proximal Land Use and Network Characteristics on Link Travel Time

The focus of this research is to investigate the influence of land use and network characteristics within the proximal area of a road link on travel time. Data for 259 links within the city of Charlotte, North Carolina, were used for the analysis. Thirty-five different types of land use characteristics were considered in this research. The proximal spatial dependency was examined by considering and capturing the land use characteristics within 0.805 km (0.5 mi), 1.61 km (1 mi), 3.22 km (2 mi), and 4.83 km (3 mi) of the selected links. Network characteristics of the upstream, downstream, cross-street, and intersecting links were also considered to incorporate the spatial dependency. Models were developed by including all the predictor variables (backward elimination) and also by selecting the predictor variables that are not correlated to each other based on the Pearson correlation coefficients. In addition to modeling using data for all the selected links, the influence of land use and network characteristics on travel time was examined by classifying the links based on the area type [central business district (CBD), CBD fringe/other business district (OBD), and urban area] and also by classifying the links based on the speed limit [56.4 or 64.4 km/h (35 or 40 mph), 72.5 or 80.5 km/h (45 or 50 mph), and >80.5 km/h (50 mph)]. Log-link with the gamma distribution model was observed to best fit the data used in this research. The results indicate that different sets of land use and network characteristics contribute to the average travel time when modeled by the buffer width. The influence of land use characteristics was observed to vary by the area type in which the link is located and the speed limit of the link. A 1-mile buffer width was observed as the best proximal area to capture land use characteristics and to estimate the average travel time at a link level. The findings help the practitioners in land use and transportation planning decisions.