06/24/2025
How do urban morphology, socio-demographics, and environmental factors influence mental health vulnerabilities in NYC?
As part of the 2024–25 CUSP Capstone Cycle, Class of 2025 graduates Swati Sharma, Qianyong Hu, and Wujun Zhou worked with Sponsor Federico Messa (Systematica, Transform Transport) and Mentors Andrea Gorrini (Transform Transport); Giulia Ceccarelli (Systematica & Transform Transport ); and Rawad Choubassi on a project titled “The Body and the City: Exploring the Complex Relationship between Mental Health and Urban Characteristics.”
Mental health is a critical urban challenge, influenced by factors such as urban morphology, socio-demographics, and environmental conditions. In dense cities like NYC, disparities in access to green spaces, housing quality, and exposure to crime can exacerbate mental health vulnerabilities. This project seeks to identify key urban characteristics influencing mental health and provide data-driven insights for urban planning interventions.
🧠 Methodology
A multi-method geospatial approach was employed to analyze the relationship between the built environment and mental health outcomes. Key methods included (1) Spatial Regression to model the relationship between urban features (e.g., land use, green spaces, and transportation) and mental health outcomes, accounting for spatial dependencies; (2) Spatial Autocorrelation using Moran’s I and Getis-Ord Gi* to detect clusters and hotspots of mental health risks across NYC; (3) Spatial Principal Component Analysis (PCA) to reduce dimensionality and identify dominant spatial patterns in urban features; (4) Predictive Modeling with logistic regression and random forest to predict mental health outcomes based on built environment characteristics like walkability, green space proximity, and housing density; and (5) Street View CNN Model using Google Street View imagery and a convolutional neural network (CNN) trained with ZenSvi to extract variables like urban degeneration and environmental quality. The analysis integrates geospatial data with advanced statistical, machine learning, and deep learning techniques, leveraging tools such as Python (GeoPandas, scikit-learn, PySAL, TensorFlow), ArcGIS, and QGIS. The findings inform urban policy to enhance mental health resilience and equity, providing actionable insights to improve NYC’s built environment for psychological well-being.
🗺️ Deliverables: Processed Raw and Clean Dataset, ArcGIS StoryMap, Data Visualization Dashboard
Interested in collaborating with CUSP through the Capstone Program? Submit your proposal by July 11, 2025: https://ow.ly/U7k450WejnZ