The value of route planning data for origin-destination estimation and prediction

Abstract

The origin-destination (OD) estimation and prediction have received lots of attention over the last several decades. However, few approaches considered route planning data. The data come from mobile map applications where some travelers plan their trips before departure. To some extents, they imply future OD demand. Motivated by this, the paper compares the volumes of planned trips and actually occurred trips extracted from trajectory data. Strong correlation (more than 0.97) exists between them. Planned trips are generally more than actual trips, especially for long-distance travel (64% more). For morning and evening peak hours, their volumes fluctuate more. These conclusions show the value of route planning data for OD estimation and prediction. Specific approaches can be studied on the basis of this preliminary analysis, but they are not included in this paper.