Food Deserts or Food Oases? Predicting Grocery Store Locations in Hamilton, Ontario

Hamilton
Grocery store
Linear regression
R package
This is the final project for the course GEOG 712 Reproducible Research Workflow with GitHub and R at McMaster University in Fall 2024 supervised by Dr. Antonio Paez.
Author

Zehui Yin

Published

December 6, 2024

Abstract:

Grocery stores play a crucial role, especially for urban residents, as they provide essential daily food supplies. The locations of grocery stores are not randomly chosen but are the result of detailed decision-making processes by grocery companies. Understanding the locations these grocers choose to establish themselves is important for public health and urban planning, as geographic access to grocery stores impacts personal health. In this paper, I utilize open data to examine grocery store locations in Hamilton, Ontario, as a case study. A zero-inflated negative binomial regression model with spatial lagged terms is fitted and estimated using maximum likelihood methods. I identified noticeable spatial patterns in grocery store locations. Grocery stores tend to cluster in nearby dissemination areas, but when there are too many grocery stores, they tend to disperse. The number of grocery stores is also significantly associated with population density, dissemination area size, the percentage of residents who do not speak an official language at home, those living in single detached houses, and the distance to Hamilton downtown.

This project addresses the following research question:

You can find the code used in this project by clicking here, and the data used in this project packaged into an R package here.

If the embedded PDF is not displayed properly, please click here.