Investigating Sprawl using AIC and Recursive Partitioning Trees: A Machine Learning Approach to Assessing the Association between Poverty and Commute Time

TL;DR

This paper uses machine learning techniques, specifically AIC and recursive partitioning trees, to analyze the relationship between urban sprawl, poverty, and commute times in the U.S., highlighting sprawl's social impacts.

Contribution

It introduces a novel application of AIC and recursive partitioning trees to assess the association between poverty and commute time related to urban sprawl.

Findings

Sprawl is linked to increased commute times.

Poverty correlates with areas affected by sprawl.

Machine learning effectively models urban socio-economic patterns.

Abstract

Sprawl, according to Glaeser and Kahn, is the 21st century phenomenon that some people are not dependent on city-living due to automobiles and therefore can live outside public transportation spheres and cities. This is usually seen as pleasant and accompanied by improved qualities of life, but as they addressed, the problem remains that sprawl causes loss of jobs for those who cannot afford luxurious alternatives but only inferior substitutes (Glaeser and Kahn 2004). Therefore, through our question, we hope to suggest that sprawl has occurred in the U.S. and poverty is one of the consequences.