My Portfolio
Research archive
Environmental Economics & Choice ModellingPrimary-data study · Research in Progress

Last-Mile Delivery Preferences under Urban Flood Risk

Poster summary

Discrete choice experiment · conditional logit · mixed logit · latent class models · scenario simulation · survey data engineering

A primary-data discrete choice experiment on how adverse weather and urban flooding shape trade-offs across last-mile delivery options.

01

Research question

How do urban flood conditions change consumer trade-offs among delivery time, cost, reliability, flood severity, and related service attributes?

02

Methodology

Repeated discrete choice tasks with approximately four to five experimentally varied attributes; planned conditional logit, mixed logit, preference-heterogeneity analysis, and latent classes if sample size and data quality permit

03

Dataset

Primary survey data from a Hanoi urban context, supported by purpose-built frontend and backend survey infrastructure, a database, and data-quality controls

04

Main finding

Primary-data collection and econometric analysis are in progress; no claim of representativeness for Vietnam or estimated population preferences is made at this stage.

Last-Mile Delivery Preferences under Urban Flood Risk

Overview

This project investigates how urban flooding and adverse weather alter preferences over last-mile delivery services. It is designed as a primary-data research study rather than a general-attitude survey.

Research Question

How do consumers trade off delivery time, delivery cost, reliability, flood severity, and related service characteristics when weather conditions worsen?

Research Design

Respondents complete repeated choice tasks in which approximately four to five delivery and flood-condition attributes vary across multiple levels. The design supports estimation of service trade-offs and preference heterogeneity in an urban Hanoi context.

Planned Analysis

The core analysis will use conditional logit and mixed logit models. Latent-class segmentation and scenario simulation are planned only if the achieved sample size and data quality support those extensions.

Research Infrastructure

The survey uses dedicated frontend and backend infrastructure, a database, and data-quality controls. These systems support the end-to-end research design; they are not presented as a standalone software project.

Limitations and Status

Sampling and fieldwork determine the appropriate scope of inference. The project does not claim national representativeness, and model extensions will remain conditional on the final sample and observed response quality.