Exploring nonlinear and interactive associations between built environment features and subjective streetscape perceptions – Nature

Report on the Relationship Between Urban Built Environments and Subjective Streetscape Perceptions in the Context of Sustainable Development Goals

1.0 Introduction: Aligning Urban Perceptions with Sustainable Development Goals

The quality of urban streetscapes is a critical determinant in achieving several Sustainable Development Goals (SDGs), particularly SDG 11 (Sustainable Cities and Communities) and SDG 3 (Good Health and Well-being). Streetscapes function as more than transportation corridors; they are fundamental to the public realm, influencing social cohesion, mental health, perceived safety, and economic vitality. This report details a study conducted in Jeonju, South Korea—a city experiencing urban shrinkage—to analyze the complex relationships between built environment features and subjective human perceptions of streetscapes. By leveraging artificial intelligence (AI) and interpretable machine learning, the study provides insights into how urban planning can create more inclusive, safe, resilient, and sustainable environments. The research specifically examines six perceptual attributes: beauty, wealth, safety, liveliness, boredom, and depression, linking them to urban form characteristics indicative of demographic and economic shifts. This approach contributes to a deeper understanding of how to revitalize urban areas in line with the human-centered principles of the SDGs.

2.0 Methodology: An AI-Driven Framework for SDG-Informed Urban Analysis

To quantify and analyze subjective urban experiences, a two-stage methodological framework was employed, demonstrating an innovative approach aligned with SDG 17 (Partnerships for the Goals) by integrating advanced technology with urban planning research.

1. Perceptual Score Estimation: A Convolutional Neural Network (CNN) was trained on the crowdsourced Place Pulse 2.0 dataset. This model was then applied to 74,550 Street View Images from Jeonju to estimate scores for the six perceptual dimensions (beauty, wealth, safety, liveliness, boredom, and depression).
2. Relationship Analysis: A Gradient Boosting Regressor model, combined with interpretable machine learning techniques, was used to examine the nonlinear and interactive associations between the estimated perception scores and a comprehensive set of built environment variables. These variables were categorized to reflect core mechanisms of urban shrinkage and their relevance to sustainable development:
   • Economic Vitality (relevant to SDG 8: Decent Work and Economic Growth)
   • Urban Design Quality (relevant to SDG 11.7: Access to Public Spaces)
   • Socio-Demographic Conditions (relevant to SDG 10: Reduced Inequalities)
   • Transportation Accessibility (relevant to SDG 11.2: Affordable and Sustainable Transport Systems)

3.0 Key Findings: Built Environment Impacts on Urban Sustainability and Well-being

The analysis revealed several significant nonlinear and interactive relationships between the built environment and streetscape perceptions, with direct implications for various SDG targets.

3.1 Feature Importance and Correlational Patterns

• Public Transport Accessibility: This emerged as the most influential factor across all perceptual dimensions, underscoring the critical role of sustainable transport infrastructure (SDG 11.2) in shaping positive urban experiences.
• Economic and Aesthetic Perceptions: A positive correlation was found between perceptions of ‘Beautiful’ and ‘Wealthy’ (r=0.141), while ‘Wealthy’ was negatively correlated with ‘Depressing’ (r=–0.138). This suggests that economic prosperity, a goal of SDG 8, is visually and emotionally linked to the quality of the urban environment.
• Safety and Liveliness: A negative correlation between ‘Lively’ and ‘Safe’ (r=–0.385) was observed, challenging traditional urban planning theories. This finding highlights the need for nuanced design strategies to ensure that vibrant urban spaces also support SDG 11.7 by being safe and inclusive for all.

3.2 Nonlinear Effects on Perceptions

The study identified key thresholds where the impact of built environment features on perceptions changes significantly.

1. Housing Prices and Perceptions:
   • Perceptions of beauty and safety generally increase with housing prices. However, the positive association with safety marginally reverses in neighborhoods with exceptionally high property values. This complex relationship is relevant to SDG 10 and SDG 11.1 (Safe and Affordable Housing), indicating that economic value alone does not guarantee perceived security.
   • Feelings of depression were higher in areas with high housing prices but low business density, pointing to a disconnect between residential value and the availability of local services and amenities, a key aspect of community well-being (SDG 3).
2. Building Age and Heritage:
   • Beauty scores peaked in areas with a 50–70% proportion of older buildings, suggesting an appreciation for architectural heritage. This supports SDG 11.4 (Protect Cultural and Natural Heritage), as well-maintained historic areas contribute positively to urban aesthetics. However, perceptions declined when the proportion exceeded 80%, likely indicating physical deterioration.
3. Urban Density and Activity:
   • Neighborhoods with simultaneously low population and employment densities were strongly associated with higher boredom scores. This finding emphasizes that compact, mixed-use development, a principle of sustainable urbanization (SDG 11.3), is crucial for creating lively and engaging environments that support mental well-being (SDG 3).

3.3 Interaction Effects

The interplay between different environmental features was found to be a powerful driver of perception.

• Beauty: The highest beauty scores were found in areas characterized by both a high proportion of older buildings and elevated housing prices, reinforcing the economic and aesthetic value of preserving cultural heritage (SDG 11.4).
• Wealth: Areas with high population density and close proximity to bus stops were perceived as wealthier, highlighting the link between sustainable transport access (SDG 11.2) and perceived economic status.
• Boredom and Depression: Low-density areas lacking both residents and jobs were perceived as more boring. Furthermore, neighborhoods with few old buildings and low business density scored higher on depression, indicating that a lack of historic character and commercial activity can negatively impact mental health (SDG 3).

4.0 Discussion and Policy Implications for Sustainable Urban Development

The study’s findings offer critical insights for urban planners and policymakers, particularly in shrinking cities, striving to meet the Sustainable Development Goals.

• Human-Centered Planning for SDG 11: The research demonstrates that subjective perceptions are a vital metric for assessing urban quality. To build sustainable communities, planning must move beyond purely economic or demographic indicators and incorporate how residents experience their environment. This aligns with the call for inclusive and people-centered urbanization in SDG 11.3.
• Prioritizing Sustainable Mobility: The strong influence of public transport accessibility on positive perceptions reinforces the need to invest in and maintain robust transit networks. Such investments are crucial for achieving SDG 11.2 and enhancing overall urban livability.
• Integrating Heritage and Revitalization: The positive perception of well-maintained older buildings suggests that integrating cultural preservation (SDG 11.4) into urban regeneration strategies can enhance aesthetic appeal and foster a stronger sense of place, contributing to community well-being (SDG 3).
• Combating Urban Decline: For shrinking cities, policies should focus on fostering moderate density and mixed-use development to counteract feelings of boredom and depression. Supporting local businesses and maintaining vibrant public spaces can improve psychological resilience and create more sustainable and equitable communities (SDG 10, SDG 11).

5.0 Conclusion

This report confirms that the physical characteristics of the built environment have profound, complex, and interactive effects on the subjective perceptions of urban residents. By employing an AI-driven analytical approach, this study provides empirical evidence that links urban design and planning decisions directly to outcomes relevant to the Sustainable Development Goals. The findings underscore that creating beautiful, safe, and lively streetscapes is not merely an aesthetic endeavor but a fundamental strategy for fostering good health and well-being (SDG 3), promoting inclusive economic activity (SDG 8), and building sustainable cities and communities (SDG 11). For cities facing urban shrinkage, these insights offer a roadmap for human-centered revitalization that can enhance the quality of life and promote long-term resilience.

Analysis of Sustainable Development Goals in the Article

1. Which SDGs are addressed or connected to the issues highlighted in the article?

The article on streetscape perceptions and the built environment in Jeonju, South Korea, connects to several Sustainable Development Goals (SDGs) through its focus on urban planning, quality of life, economic factors, and technological innovation. The primary SDGs addressed are:

• SDG 11: Sustainable Cities and Communities: This is the most central SDG to the article. The study directly examines the quality of the urban fabric, focusing on streetscapes as fundamental components of cities. It analyzes factors like perceived safety, accessibility to public transport, the role of cultural heritage (older buildings), and the overall subjective experience of urban residents, which are all core components of making cities inclusive, safe, resilient, and sustainable. The research aims to inform urban revitalization and human-centered planning, particularly in the context of “urban shrinkage.”
• SDG 3: Good Health and Well-being: The article explicitly links the built environment to mental health. It measures subjective perceptions of “depression” and “boredom” and analyzes how factors like population density and commercial activity influence these negative feelings. The introduction states that streetscapes influence “mental health outcomes,” directly aligning with SDG 3’s goal of promoting mental health and well-being.
• SDG 8: Decent Work and Economic Growth: The study incorporates several economic indicators to understand streetscape perceptions. It analyzes the impact of “housing prices,” “employment density,” and “business density” on how a street is perceived (e.g., wealthy, lively, or boring). The discussion of “economic disinvestment” and “urban vitality” in shrinking cities connects the physical environment to local economic health and productivity.
• SDG 9: Industry, Innovation, and Infrastructure: A significant contribution of the study is its methodology. It highlights the “utility of AI-based approaches in the field of urban planning” by using advanced technologies like Convolutional Neural Networks (CNN) and interpretable machine learning (IML). This demonstrates the application of innovation and scientific research to better understand and plan urban infrastructure, such as public transport systems, whose accessibility was found to be a critical factor.

2. What specific targets under those SDGs can be identified based on the article’s content?

Based on the article’s detailed analysis, several specific SDG targets can be identified:

1. Target 11.2: “By 2030, provide access to safe, affordable, accessible and sustainable transport systems for all…” The article identifies “public transport accessibility” as the “most influential variable across all dimensions” of perception. It specifically measures proximity to bus stops and analyzes its relationship with perceptions of safety and wealth, directly addressing the importance of accessible transport systems.
2. Target 11.3: “By 2030, enhance inclusive and sustainable urbanization and capacity for participatory, integrated and sustainable human settlement planning and management…” The study’s stated goal is to provide “policy-relevant insights for urban revitalization efforts” and inform “more human-centered planning processes.” By analyzing how residents perceive their environment, the research provides a basis for more inclusive and sustainable urban planning that considers subjective experience.
3. Target 11.4: “Strengthen efforts to protect and safeguard the world’s cultural and natural heritage.” The article finds that a higher “proportion of old buildings” is associated with perceptions of beauty and liveliness. This suggests that preserving architectural heritage and historic character is crucial for enhancing the perceptual quality of urban spaces, aligning with this target.
4. Target 11.7: “By 2030, provide universal access to safe, inclusive and accessible, green and public spaces…” The entire study revolves around the quality of public spaces (streetscapes). It directly measures “perceived safety” and analyzes how factors like proximity to the downtown core and transit infrastructure contribute to it, which is central to this target.
5. Target 3.4: “By 2030, reduce by one third premature mortality from non-communicable diseases through prevention and treatment and promote mental health and well-being.” The research quantifies perceptions of “depression” and “boredom” and links them to specific built environment features, such as low population and employment densities. This provides empirical evidence on how urban design can impact mental well-being.
6. Target 9.5: “Enhance scientific research, upgrade the technological capabilities of industrial sectors in all countries…” The article’s methodology, which employs “a Convolutional Neural Network (CNN),” “Gradient Boosting Decision Tree (GB) model,” and “interpretable machine learning (IML) techniques,” is a direct application of advanced technology to enhance scientific research in the field of urban planning.

3. Are there any indicators mentioned or implied in the article that can be used to measure progress towards the identified targets?

Yes, the article mentions and uses several quantitative and qualitative metrics that can serve as indicators for measuring progress towards the identified targets.

• Perceptual Scores as Indicators: The primary indicators developed in the study are the six subjective perception scores: beauty, wealth, safety, liveliness, boredom, and depression. These scores, estimated via a CNN model, serve as direct indicators for:
  • Target 11.7 (perceived safety and inclusivity of public spaces).
  • Target 3.4 (mental well-being, measured through scores for depression and boredom).
• Built Environment Indicators: The independent variables used in the analysis are concrete indicators related to urban form and function.
  • For Target 11.2 (transport accessibility), the indicator is distance to bus stops.
  • For Target 11.4 (cultural heritage), the indicator is the proportion of old buildings in a neighborhood.
  • For Target 11.3 (sustainable urbanization), indicators include population density, employment density, business density, and building density.
  • For Target 8.5 (economic vitality), indicators include average housing price, employment density, and business density.
• Technological Application as an Indicator: For Target 9.5 (innovation), the use of specific AI-based methodologies (CNN, IML) for urban analysis is itself an indicator of the integration of advanced technology in scientific research for sustainable development.

4. Table of SDGs, Targets, and Indicators

Source: nature.com