Generating online freight delivery demand during COVID-19 using limited data

Urban freight data analysis is crucial for informed decision-making, resource allocation, and optimizing routes, leading to efficient and sustainable freight operations in cities. Driven in part by the COVID-19 pandemic, the pace of online purchases for at-home delivery has accelerated significantly. However, responding to this development has been challenging given the lack of public data. The existing data may be infrequent because of survey participant non-responses. This data paucity renders conventional predictive models unreliable.

We address this shortcoming by developing algorithms for data imputation and replication for future urban freight demand given limited ground truth online freight delivery data. Our generic framework is capable of taking in repeated cross-sectional surveys and replicating frequent samples from them. In this paper, our case study is focused on Puget Sound Regional Council (PSRC) household travel survey (HTS) data restricted to the Seattle–Tacoma–Bellevue, WA Metropolitan Statistical Area (MSA). We show how to impute the missing online freight deliveries in our travel survey dataset from ground truth values by making a similarity-based matching between the samples of missing and available online delivery volumes. Empirical and theoretical analyses both demonstrate high veracity of imputation where the estimated freight delivery volumes closely resemble the ground truth values. Utilizing the obtained similarity-based matching, we show how to generate data across future and past travel survey datasets with an emphasis on maintaining some consistent trends across the datasets. This work furthers existing methods in demand estimation for goods deliveries by maximizing available scarce data to generate reasonable estimates that could facilitate policy decisions.