Urban sprawl is the growing expansion of cities into undeveloped land, an ongoing issue for cities worldwide. As urban populations grow, more land is required to accommodate people and the associated economic development required to support that growth. This expansion is generally not planned for; it happens in large part organically. Without proper planning to outline roadway infrastructure, essential services, and open space, the result can lead to overcrowding, unaffordable housing, and congestion. A promising solution is to employ machine learning models. Their ability to ingest data, solve complex problems, and uncover hidden patterns using diverse data sources (including satellite, drone, terrestrial transport, and mobile devices) will be key to better urban planning.
In “Application of Machine Learning Algorithms to Predict Urban Expansion” for the Journal of Urban Planning and Development, authors Rejira K. Robi and Jain K. George explore the complexity and variables of urban expansion and how technological advances can improve urban development. A literature review helped them identify which models and algorithms were most likely to aid in their research, and they accounted for various factors, including complexity, scalability, reliability, efficiency, readability, modularity, and testability. Their findings indicate that using deep learning models will help communities make more informed decisions, optimize resource allocation, and devise sustainable urban development strategies. Learn more about how predictive models will help urban planners and policymakers to manage urban expansion at https://ascelibrary.org/doi/10.1061/JUPDDM.UPENG-5466. The abstract is below.
Abstract
Urban expansion presents significant challenges for sustainable development. Predicting urban growth patterns is crucial for effective urban planning and resource management. This review explores the application of machine learning models in predicting urban expansion. This study aims to systematically review existing literature on using machine learning models for urban expansion prediction. We hypothesize that machine learning, particularly deep learning techniques, can offer valuable insights and improve the accuracy of urban growth predictions compared to traditional methods. A comprehensive literature search was conducted using relevant databases to identify research articles addressing urban expansion prediction with machine learning models. The search strategy included keywords related to urban growth, expansion, machine learning, and various model types. Inclusion and exclusion criteria were established to ensure the relevance and quality of the retrieved studies. Data extraction focused on the types of urban growth models, variables considered, machine learning methodologies employed, and the effectiveness of the models in predicting urban expansion. The review identified various types of machine learning models used for urban expansion prediction, including shallow learning (e.g., random forest, support vector machines) and deep learning (e.g., convolutional neural networks, long short-term memory) architectures. Studies analyzed the influence of diverse variables such as infrastructure, demographics, and environmental factors on urban growth patterns. The review found that deep learning models generally demonstrated superior performance to shallow learning models in predicting urban expansion due to their ability to handle complex spatial data and relationships. Case studies from China and Korea provided practical examples of applying these models and showcased their potential for real-world urban planning applications. This review highlights the growing potential of machine learning models, particularly deep learning, for predicting urban expansion. By incorporating various urban growth indicators and leveraging advanced learning algorithms, these models offer a data-driven approach for informed urban planning decisions. Further research is needed to explore the integration of explainability techniques within these models and their application in diverse geographic contexts.
Explore more about the potential for machine learning algorithms to help urban planners improve zoning and planning in the ASCE Library: https://ascelibrary.org/doi/10.1061/JUPDDM.UPENG-5466.
