- 场景营城的理念与实践
- 文章编号：1009-6000(2022)10-0032-08
- 中图分类号：F281 文献标识码：B
- Doi：10.3969/j.issn.1009-6000.2022.10.005
- 项目基金：国家自然科学基金面上项目“城市收缩治理的理论模型、国际比较和关键规划领域研究”（52078197）。
- 作者简介：周恺，湖南大学建筑与规划学院城乡规划系，副教授；熊益群，通信作者，湖南大学建筑与规划学院城乡规划系，硕士研究生。
- 基于无监督聚类方法的城市消费场景识别研究：以长沙为例
- A Method for Identifying Urban Consumption-Scenescapes Using Unsupervised Clustering Algorithms: A Case Study of Changsha
- 周恺 熊益群
- ZHOU Kai XIONG Yiqun
- 摘要：
新芝加哥学派的场景理论近年在规划学界和业界产生了较大影响。“场景营城”成为城市更新和活力再造的新思路，为实现城市高质量发展，提供了切合居民日常生活和需求习惯的消费空间设计方法。文章以长沙为案例，探索并实验一种基于无监督聚类方法的城市消费场景识别技术。研究利用窗口滑动方法构建场景识别单元，然后用无监督聚类方法对消费场景进行分类，最后，基于结果探究各类场景的分布规律和机制。研究发现，长沙的消费场景总体可以分为7 类，不同场景类别存在一定空间分布规律，并且互相之间存在影响。本研究将机器学习无监督聚类方法应用于场景识别的实证分析，探索了一套智能化方法，帮助规划师、管理者等有针对性地精细化识别、布局相关舒适物设施，为场景营城相关实践提供技术工具。 - 关键词：
场景理论；消费空间；无监督聚类；场景识别；大数据 - Abstract: The ¡°scene theory¡± of the New Chicago School has exerted great infl uence in urban planning in recent years. The idea of ¡°Building Urban Scenescapes¡± was proposed as an innovative perspective for urban regeneration and revitalization, which became not only a new design philosophy of commercial spaces that stimulates consumption, but also a better approach to organizing high-quality urban facilities that suit residents¡¯daily life. This paper took Changsha as a case to explore and experiment with a new method for identifying urban consumption scenescapes using Point of Interest (POI) data and machine learning algorithms. Firstly, the study selected and located ¡°urban amenities¡± that constitutes ¡°scenes¡± for consumption in cities using POI data, and defi ne units for analysis using a sliding window method. Secondly, the unsupervised clustering methods (i.e. the PCA reduction and clustering analysis) were applied to classify the consumption scenes. Finally, the pattern of distribution of various ¡°scenes¡± are explained. This study applies machine learning algorithms to the empirical analysis of urban consumptionscenescapes identification and explores a set of new methods to help planners and decision-makers to recognize and organize urban amenities. It aims to provide a new tool for the practice of ¡°Building Urban Scenescapes¡±.
- Key words: scene theory; consumption space; unsupervised clustering; scenescapes; big data