AI-Driven Technology, Geospatial Analysis, Social Vulnerability, and Child Neglect

AI Neglect # In progress

Utilizing AI-Driven Technology and Geospatial Analysis to Explore Landscape Morphology, Social Vulnerability, and Child Neglect in Los Angeles

Summary:
This project integrates AI-driven computer vision, street-level imagery, and spatial epidemiology to examine how neighborhood environmental conditions and human perceptions influence child neglect risk in Los Angeles. By applying semantic segmentation and perception models to tens of thousands of street images, this research generates fine-scale, objective measures of the built environment, including greenness, lighting, land use diversity, and neighborhood surveillance. These environmental indicators, along with social vulnerability and perceptions of safety, are analyzed in relation to child neglect patterns to inform prevention efforts and urban design strategies.

Research Questions:

  • RQ1: What is the spatial distribution of built environment features and neighborhood perceptions in the study area?
  • RQ2: How do objective measures of the built environment and subjective perceptions of neighborhood conditions relate to child neglect risk?
  • RQ3: Do neighborhood perceptions moderate the relationship between the built environment and neglect?

See related content here

Short description: This project uses AI-based semantic segmentation of street-level imagery combined with spatial epidemiology to investigate how built environment features and neighborhood perceptions shape child neglect risk in Los Angeles.
Lead developer: Dr. Gia Elise Barboza-Salerno
Further reading: Please see the presentation located here
Download: Github repo
Main data source(s): Mapillary imagery, American Community Survey, Los Angeles Open Data, Area Deprivation Index
Coverage: South Central Los Angeles neighborhoods
Citation: Barboza-Salerno, G. (2025). Utilizing AI-Driven Technology and Geospatial Analysis to Explore Child Neglect Risk in Los Angeles. Preprint.

Access the ZenSVI website here for more details.