Investigating Rideshare Patterns and Passenger Distribution: A Case Study

Ridesharing can offset negative transportation effects such as congestion and related environmental impacts and is, therefore, worthy of an in-depth investigation into the dynamics that determine its success. To this end, we employed a data set of all the ridesharing trips taken in Arlington, Texas, over a 2-year period, developed a random forest model to predict the number of rideshare passengers from various origination points across different time periods, and evaluated the impact of various features on the number of passengers who chose ridesharing as their means of transportation. The random forest model performed well for this classification task, as it was adept at predicting passenger distributions. The point of origin, with the greatest number of rides originating from the area surrounding the University of Texas at Arlington, was found to have the greatest impact on the number of passengers, followed by the time of day. The analysis also identified a noticeable increase in ridership across the 2-year interval, reflecting the growing demand for rideshare services. The results of this study can help ridesharing providers improve their levels of efficiency and service by equipping them with information about how rideshare services are utilized and empowering them to make data-driven decisions. Through valuable insights into rideshare dynamics, stakeholders may better plan for resource allocation and design of pre-positioned vehicles in high-demand areas to reduce wait times and improve customer satisfaction.