Predicting and Planning for the Future: North American Truckload Transportation
Name
scm2020-sokoloff-predicting-and-planning-for-the-future-north-american-truckload-transportation-capstone.pdf
Description
Capstone research project
Size
944.59 KB
Format
Adobe PDF
Checksum (MD5)
2d7709d156930ad8d525baf0f7b33a18
Author(s)
Sokoloff, David
Zhang, Gaohui
Date Issued
July 21, 2020
Abstract
The trucking industry is crucial to the United States economy. An overwhelming majority of goods
transported across the US are moved in trucks. For most companies, truck transportation is a
prominent component that impacts their production, warehousing, customer service, and overall
business performance. In fact, trucking constitutes one of the largest operational costs for a
company. Trucking costs are highly volatile due to their association with the capricious freight
industry and the US economy. Unexpected market fluctuations inevitably disturb companies’
budget planning and operations, as well as impact their profits. This paper formulates a machine
learning model to predict the US truckload dry van spot rate and a playbook of contingent actions.
The model variables target and recognize the key elements in the trucking industry and the
economy. Tested across 6 years of data, the model achieved an average MAPE below 7% and
mean error below 0.05 for predicting 12 months in the future. The strong forecast accuracy allows
companies to employ our playbook’s strategic and tactical measures to mitigate risk and unplanned
costs stemming from the volatility in the US trucking market.
Subjects
Transportation
Persistent DSpace Link
