The primary objective of this study was to conduct a spatial analysis for the usage of ride-sourcing services. This study collected the ride-sourcing trip data from Uber in New York City, as well as the road network attributes and the socio-demographic characteristics. The collected data were further aggregated into 167 ZIP code tabulation areas (ZCTA). The geographically weighted regression models were developed to establish the relationships between ride-sourcing usage and various contributing factors on weekday and weekend, respectively. The results suggested that the number of population with bachelors’ degree or higher, the number of parking spaces, the ratio of floor area allocated for commercial use, and the subway accessibility in each ZCTA were positively correlated with the Uber usage. The number of unemployment population and the median household income in each ZCTA were found to be negatively correlated with the Uber usage. The results of the present study could provide useful insights for predicting the demand of zonal-level ride-sourcing services and developing effective policy initiatives for pricing and regulation of ride-sourcing industry.

Shared lane design has been used to relieve the hindrance of left-turn movement. The primary object of the study is to identify the benefit zone where shared lane design overwhelms conventional protected control strategies at signalized intersections. Analytical models were proposed to calculate the minimum green time duration required to fully discharge all the approaching vehicles, which is used as an indicator of the effectiveness of different strategies. Based on the proposed model, the benefit zone of the shared lane design was identified. With the benefit zone plot, engineers can decide whether to deploy shared lane design according to the traffic demand. A realistic method was also provided to increase the accuracy of the proposed models. The results showed that the shared lane design is suitable to intersections where traffic demand from opposing approaches is significantly unbalanced. The proposed framework is expected to be helpful for traffic engineers to improve the performance of signalized intersections effectively and efficiently without complex computing.