TL;DR
This study investigates how young adults perceive and interact with mountain city parks, revealing that naturalness, aesthetics, and specific spatial features like elevation changes significantly impact their satisfaction. By employing advanced statistical methods, the research provides actionable insights for designing more engaging and restorative urban green spaces in complex, hilly environments.
What Problem Does This Paper Address?
Traditional urban park designs often fall short in meeting the diverse needs of today's youth, especially in the unique context of mountainous cities. These cities, characterized by their challenging topography and integrated urban-natural environments, present distinct spatial and perceptual challenges for park users. While research has explored urban parks' impact on well-being, there's a lack of localized studies focusing on how young people experience and perceive these spaces, particularly in complex mountain city settings. This gap hinders the development of effective, human-centered design strategies for these vital urban green areas.
Methods
The researchers employed a multi-faceted approach:
- Principal Component Analysis (PCA): Used to identify the key visual and auditory indicators most strongly correlated with environmental satisfaction across different park gathering spaces. This helped reduce complex data into more manageable components.
- Random Forest (RF) Model: A machine learning algorithm used to analyze non-linear relationships between spatial elements and subjective evaluations (environmental satisfaction). This model also ranked the importance of various spatial factors.
- Shapley Additive Explanations (SHAP) Analysis: Applied to further dissect the RF model's findings, quantifying the specific impact (both magnitude and direction) of individual spatial elements on user satisfaction. This provided a deeper understanding of "why" certain features matter.
- Audiovisual Walking Surveys: Participants experienced the parks firsthand, collecting on-site questionnaire data and audio recordings to capture real-time perceptions.
Datasets
The study utilized data collected from 46 university students (aged 19-25) who participated in audiovisual walking surveys in two mountain city parks in Yuzhong District, Chongqing: Pipa Mountain Park and Eling Park. A total of 495 valid questionnaires and 495 valid audio samples were analyzed. The study also defined and measured 16 visual spatial indicators (e.g., green view index, openness) and 9 soundscape spatial indicators (e.g., water sounds, traffic noise).
Key Insights and Ideas
- Elevation is a Double-Edged Sword: While elevation changes positively contribute to satisfaction at elevated viewpoints (offering expansive vistas), they can negatively impact overall satisfaction in general park areas and other gathering spaces, likely due to accessibility challenges or the need for retaining walls.
- Naturalness and Aesthetics Reign Supreme: These two factors consistently emerged as the most significant drivers of environmental satisfaction, underscoring the fundamental human need for connection with nature.
- Context Matters for Spatial Elements: The importance of specific spatial features varies greatly depending on the type of gathering space. For instance, entrance links and green view index are crucial at entrance platforms, while water features and color types are more influential on path platforms.
- Youth Preferences Lean Towards Natural Sounds: Participants showed a preference for natural sounds like birdsongs, rustling leaves, and water, while artificial sounds (conversations, music) were less favored.
Key Results
- Strong Correlations: The first principal component of the visual environment in entrance platforms and key nodes, and the auditory environment in entrance platforms, path platforms, and elevated points, showed significant positive correlations with environmental satisfaction (p < 0.01).
- Key Influencers: Naturalness and aesthetics were identified as the primary factors influencing environmental satisfaction.
- SHAP Analysis Highlights: Green view index, cultural buildings, plant stratification, and insect chirping sounds positively impacted overall satisfaction. Conversely, elevation drop, openness, traffic sounds, and the number of entrance links had negative influences.
- Model Performance: The random forest model demonstrated high accuracy in predicting environmental satisfaction, with R² values generally above 0.6 for both training and testing sets across different gathering spaces.
Limitations and Future Work
The authors acknowledge several limitations:
- Subjectivity: Data relied on self-reported questionnaires, which can be influenced by personal emotions and cultural backgrounds.
- Sample Representativeness: The study focused on young university students, potentially not reflecting the perceptions of broader age groups or diverse user demographics. The selected parks, while representative of mountain city parks, might not encompass all variations.
- Future Directions:
- Incorporate objective measurements (e.g., physiological data) to complement subjective feedback.
- Expand the study to include a wider range of mountain city parks and diverse age groups.
- Develop a comprehensive database of environmental factors and user perceptions for computational modeling and automated park design.
- Explore the application of artificial neural networks for more robust analysis.
Summary updated