Using machine learning methods to explore the urban expansion patterns and planning topics : a case study in Texas Triangle

Le résumé fourni par la source

Urban expansion is a complex process shaped by increasing population, economic activity, and the continual conversion of land to urban use. While extensive research has examined the spatial dynamics of urban expansion, significant gaps remain in linking these patterns with planning policy at a regional scale. This dissertation addresses these gaps by investigating how local planning practices, as reflected in comprehensive plans, relate to urban expansion across the four major metropolitan areas in the Texas Triangle megaregion: Dallas, Houston, San Antonio, and Austin. The study first uses the National Land Cover Dataset (NLCD) and develops a spatial framework to categorize urban expansion types among cities in the study area. K-means clustering and spatial analysis are then applied to uncover heterogeneous urban expansion trajectories. To explore the policy dimension, the dissertation selects fifty representative cities and conducts a topic modeling analysis of their most recent comprehensive plans. Using natural language processing techniques and innovatively leveraging the ChatGPT model with human oversight, the research extracts and interprets key policy themes embedded in the plans. Finally, a statistical analysis is conducted to examine the relationship between urban expansion clusters and key topics in the comprehensive plans. The research identifies five clusters of cities, ranging from minimal growth to high expansion rates. Topic modeling reveals eighteen key policy topics across the comprehensive plans, spanning land use, sustainable development, transportation, and other planning domains. The statistical analysis further indicates variation in planning practices by city type and growth trajectory. This dissertation contributes to urban planning scholarship by advancing methodological approaches that bridge empirical spatial research with policy analysis. It demonstrates the potential of integrating geospatial data and planning documents with emerging AI tools to gain deeper insights into the relationship between urban development and local policy. The findings offer practical implications for planners and policymakers, highlighting the importance of considering place-specific growth dynamics when crafting planning strategies.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.