Weekday Travel Demand Distribution Pattern of Ride-Hailing Service Based on Didi Daily Order Data in Beijing

A city is a multi-layered open complex giant system, including graph networks of road, rail, air, and water transportation system. Focused on time-dependent mobility patterns, this study analyzed the weekday daily travel pattern of transportation network company (TNC) riders based on Didi daily order data in Beijing in May 2021. Especially, trips arrived in and departed from traffic analysis zones with rail stations and airports was visualized on to analyze the trip distribution by time of day. Citywide OD matrices at traffic analysis zone level for ride-hailing service vehicles was then derived and OD Matrix Estimation (ODME) was also conducted for OD matrices of car, bike and electric bike based on traffic counts data collected in 2019. Based OD matrices of different vehicle class, using bi-conjugate Frank-Wolfe algorithm built in Cube Voyager platform, the highway assignment procedure was processed to estimate volume by mode for links in the roadway network by different time of day. Ride hailing service users travel along major freeway corridor during peak hours to major employer and commercial centers in city proper area. Results of volume forecasting can be validated using screenline traffic count data.