E-commerce choice

Contents

E-commerce choice#

A household-level e-commerce model is developed to generate demand for delivery stops within POLARIS using data from the WholeTraveler Transportation Behavior Study by Lawrence Berkeley National Laboratory. A hurdle ordered probit model was applied to estimate households’ e-commerce activity. The model is conceptualized as a two-stage model, where, in the first stage, households’ binary choice to participate in e-commerce activity is estimated. Subject to the participation decision, the second stage of the model quantifies the ratio of delivery to retail shopping for participating households. A number of household characteristics, land-use/neighborhood characteristics, and interactions between household and land-use/neighborhood characteristics are used to estimate the model. The model is then implemented in the POLARIS modeling workflow to evaluate whether each household participates in e-commerce and, if it does, how frequently it participates (in other words, how many orders it makes per week). The weekly e-shopping deliveries then replace physical shopping trips in the model. Each POLARIS model (e.g., Detroit, Atlanta, Austin, etc.) should then be calibrated for the corresponding region.

References#

  • Stinson, M., Enam, A., Moore, A., & Auld, J. (2019, September). Citywide impacts of E-commerce: Does parcel delivery travel outweigh household shopping travel reductions?. In Proceedings of the 2nd ACM/EIGSCC Symposium on Smart Cities and Communities (pp. 1-7). https://dl.acm.org/doi/pdf/10.1145/3357492.3358633

  • Khan, N.A., Sahin, O., Uhm, H., & Auld, J. (2026). Modeling the Digital Economy’s Transportation Footprint: An Integrated Analysis of Telecommuting, E-commerce, and Land Use using POLARIS Transportation Systems Simulator. Transportation Research Record (in-preparation).

Contents