A data-informed analytical approach to human-scale greenway planning: Integrating multi-sourced urban data with machine learning algorithms

《A data-informed analytical approach to human-scale greenway planning: Integrating multi-sourced urban data with machine learning algorithms》

打印

关键字: Greenway planning;Multi-sourced urban data;Data-informed urban design;Machine learning algorithms

作者单位: College of Architecture and Urban Planning, Tongji University, China;Joint International Research Laboratory of Eco-Urban Design, Tongji University, China;Shanghai Tongji Urban Planning and Design Institute, China;College of Architecture and Urban Planning, Tongji University, China;Joint International Research Laboratory of Eco-Urban Design, Tongji University, China;Shanghai Tongji Urban Planning and Design Institute, China

摘要: Urban greenways have been recognized as an important strategy to improve human-scale quality in high-density built environments. Nevertheless, current greenway suitability analysis mainly focuses on geographical and natural issues, failing to account for human-scale urban design factors. Accordingly, this study proposes a data-informed approach to planning urban greenway networks using a combination of classical urban design theories, multi-sourced urban data, and machine learning algorithms. Maoming City in China was used as a case study. Per classical urban design theories, specifically, Cervero and Ewing’s 5D variables, density, diversity, design, dimensions of destination accessibility, and distance-to-transit, were selected as key factors. A series of new urban data, including points of interest (PoIs), location-based service (LBS) positioning data, and street view images, were applied in conjunction with machine learning algorithms and geographical information system (GIS) tools to measure these key factors at a human-scale resolution and generate an optimized greenway suitability analysis. This analytical approach is an attempt to take human-scale concerns into account on a city-wide scale regarding greenway network generation. It also pushes the methodological boundaries of greenway planning by combining classical urban design thinking with new urban data and new techniques.