| Description |
Data-driven analytics has been successfully utilized in many experience-oriented areas, such as education, business, and medicine. With the profusion of traffic-related data from Internet of Things and development of data mining techniques, data-driven analytics is becoming increasingly popular in the transportation industry. The objective of this research is to explore the application of data-driven analytics in transportation research to improve the traffic management and operations. Three problems in the respective areas of transportation planning, traffic operation, and maintenance management have been addressed in this research, including exploring the impact of dynamic ridesharing system in a multimodal network, quantifying non-recurrent congestion impact on freeway corridors, and developing infrastructure sampling method for efficient maintenance activities. First, the impact of dynamic ridesharing in a multimodal network is studied with agent-based modeling. The competing mechanism between dynamic ridesharing system and public transit is analyzed. The model simulates the interaction between travelers and the environment and emulates travelers' decision making process with the presence of competing modes. The model is applicable to networks with varying demographics. Second, a systematic approach is proposed to quantify Incident-Induced Delay on freeway corridors. There are two particular highlights in the study of non-recurrent congestion quantification: secondary incident identification and K-Nearest Neighbor pattern matching. The proposed methodology is easily transferable to any traffic operation system that has access to sensor data at a corridor level. Lastly, a high-dimensional clustering-based stratified sampling method is developed for infrastructure sampling. The stratification process consists of two components: current condition estimation and high-dimensional cluster analysis. High-dimensional cluster analysis employs Locality-Sensitive Hashing algorithm and spectral sampling. The proposed method is a potentially useful tool for agencies to effectively conduct infrastructure inspection and can be easily adopted for choosing samples containing multiple features. These three examples showcase the application of data-driven analytics in transportation research, which can potentially transform the traffic management mindset into a model of data-driven, sensing, and smart urban systems. The analytica |
