Measuring Physical Disorder in Urban Street Spaces: A Large-Scale Analysis Using Street View Images and Deep Learning

Abstract

Physical disorder is associated with negative outcomes in economic performance, public health, and social stability, such as the depreciation of property, mental stress, fear, and crime. A limited but growing body of literature considers physical disorder in urban space, especially the topic of identifying physical disorder at a fine scale. There is currently no effective and replicable way of measuring physical disorder at a fine scale for a large area with low cost, however. To fill the gap, this article proposes an approach that takes advantage of the massive volume of street view images as input data for virtual audits and uses a deep learning model to quantitatively measure the physical disorder of urban street spaces. The results of implementing this approach with more than 700,000 streets in Chinese cities—which, to our knowledge, is the first attempt globally to quantify the physical disorder in such large urban areas—validate the effectiveness and efficiency of the approach. Through this large-scale empirical analysis in China, this article makes several theoretical contributions. First, we expand the factors of physical disorder, which were previously neglected in U.S. studies. Second, we find that urban physical disorder presents three typical spatial distributions—scattered, diffused, and linear concentrated patterns—which provide references for revealing the development trends of physical disorder and making spatial interventions. Finally, our regression analysis between physical disorder and street characteristics identified the factors that could affect physical disorder and thus enriched the theoretical underpinnings.

城市空间失序与经济状况、公共健康和社会稳定的负面结果相关, 如财产贬值、精神压力、恐惧和犯罪。目前, 对城市空间失序、尤其是在精细尺度上识别城市空间失序的研究较少但有增多的趋势。然而, 现在还没有一种有效、可复制和低成本的方法来评估大范围的城市空间失序。为了填补这一空白, 本文以海量街景图像作为虚拟核查的输入数据, 利用深度学习模型定量评估城市街道空间失序。我们在中国70多万条城市街道上实现了这一方法, 结果证明了该方法的有效性和效率。据我们所知, 这是全球首次尝试量化如此大规模城市区域的空间失序。通过在中国的大规模实证分析, 本文在理论上有如下几点贡献。首先, 我们拓展了以前对美国的研究中被忽视的城市空间失序因素。其次, 我们发现, 城市空间失序呈现出三种典型的空间分布：分散型、扩散型和线性集中型, 这为揭示城市空间失序的发展趋势和空间干预提供了参考。最后, 对城市空间失序和街道特征之间的回归分析, 确定了城市空间失序的可能影响因素, 从而丰富了理论基础。

El desorden físico está asociado con resultados negativos en el desempeño económico, la salud pública y la estabilidad social, como la depreciación de la propiedad, el estrés mental, el miedo y el crimen. Un cuerpo de literatura limitado, aunque en expansión, considera el desorden físico en el espacio urbano, especialmente en lo que concierne al tópico de identificar el desorden físico a escala fina. No obstante, actualmente no se dispone de un modo efectivo y replicable de medir el desorden físico a escala fina, para un área grande y a bajo costo. Para llenar este vacío, en este artículo se propone un enfoque que aprovecha el volumen masivo de imágenes con vistas a la calle como datos de input para las auditorías virtuales, y se usa un modelo de aprendizaje a profundidad para medir cuantitativamente el desorden físico de los espacios urbanos de las calles. Los resultados obtenidos con la implementación de este enfoque en más de 700.000 calles de ciudades chinas –que, para nuestro conocimiento, globalmente es el primer intento de cuantificar el desorden físico, en áreas urbanas grandes– validan la efectividad y eficiencia del enfoque. A través de este análisis empírico a gran escala acometido en China, este artículo hace varias contribuciones teóricas. En primer lugar, ampliamos el número de factores del desorden físico, anteriormente menospreciados en estudios realizados en Estados Unidos. En segundo término, descubrimos que el desorden físico urbano presenta tres distribuciones espaciales típicas –dispersión, difusión y patrones lineales concentrados– que proveen referencias para revelar las tendencias del desarrollo del desorden físico, y para realizar intervenciones espaciales. Finalmente, nuestro análisis de regresión entre el desorden físico y las características de las calles identificó los factores que podrían afectar al desorden físico, enriqueciendo de ese modo las fundamentaciones teóricas.

Key Words:

• deep learning
• physical disorder
• street view image
• urban street space

关键词:

• 深度学习
• 城市空间失序
• 街景图像
• 城市街道空间。

Palabras clave:

• aprendizaje profundo
• desorden físico
• espacio urbano callejero
• imagen de las calles

Acknowledgment

The data and code of this study are available at Mendeley Data (http://dx.doi.org/10.17632/d3d4h5bvss.1).

Supplemental Material

Supplemental data for this article can be accessed on the publisher’s site at: https://doi.org/10.1080/24694452.2022.2114417

The auditing guidance of this study is available as supplemental material. It will help readers to understand how the auditors identify the physical disorder in the process of virtual auditing. It also provides guidance for the research community in replicating the approach in future studies on physical disorder in other cities.

Notes

1 The prefecture is the basic administrative unit between province and county in China, and there are 293 prefectural cities in total until 2021 (National Bureau of Statics of China; https://data.stats.gov.cn/easyquery.htm?cn=C01). Therefore, these cities represent the most densely populated and rapidly developing urbanization regions in China.

Additional information

Funding

This work was supported by the National Natural Science Foundation of China [No. 52178044, 71834005 and No. 51878052].

Notes on contributors

Jingjia Chen

JINGJIA CHEN has earned both a master’s and a bachelor’s degree from the Department of Urban Planning, Tsinghua University, Beijing, China, 100084. E-mail:[email protected]. Her research interests include urban big data analysis and urban design for future cities.

Long Chen

LONG CHEN is a Lecturer in the Faculty of Architecture, Civil and Transportation Engineering, Beijing University of Technology, Beijing, China, 100124. E-mail: [email protected]. His research interests include transportation big data analysis and sustainable transportation.

Yan Li

YAN LI is a Research Fellow in the Department of Urban Planning at Tsinghua University, Beijing, China 100084. E-mail: [email protected]. Her research interests include geographic information systems, human dynamics, and urban science.

Wenjia Zhang

WENJIA ZHANG is an Assistant Professor in the School of Urban Planning & Design, Peking University Shenzhen Graduate School, Shenzhen, Guangdong, China, 518055. E-mail: [email protected]. His research interests include network complexity in urban and geography studies, network and spatial economics, land use and transportation planning, as well as big data and machine learning approaches.

Ying Long

YING LONG is an Associate Professor in the Department of Urban Planning, Tsinghua University, Beijing, China, 100084. E-mail: [email protected]. His research interests include (new) urban science, urban big data analysis, urban modeling, data enhanced design, smart cities, and future cities.