Synthetic Data Generation for Small-Area Demand Forecasting of Freight Flows

Paul Metaxatos
University of Illinois at Chicago, USA

Small area statistics have become increasingly critical for the planning and management of intermodal transportation systems. However, for reasons associated with disclosure of confidential information, data is often released on a fairly coarse geography vis-à-vis a much finer geographical level. This has led to extensive research on small area estimation - i.e., estimation at a more detailed geographical level based on data at a coarser level. Most of this work has been single-area-specific or non-flow data. Freight flows, at a minimum, have origin and destination location specificity, which leads to greater complexity. This paper addresses this issue providing a methodology for small-area estimation of freight flows based on the gravity model. Preliminary empirical findings using publicly available data demonstrate the reasonableness of the method as a freight-planning tool.

Download Full Paper