Equicity AI — Urban Infrastructure Risk Assessment
A GeoAI platform leveraging Google Earth Engine and machine learning to assess risk across urban infrastructure assets, with an interactive web interface for spatial risk visualization.
Interactive map — scroll and click to explore.
Context
Urban utilities and infrastructure assets face increasing climate and operational risks, yet most cities lack spatially explicit risk assessment tools. Integrating satellite data with ground-truth asset information can enable proactive risk management.
Problem Statement
How can satellite imagery and geospatial machine learning be combined to produce asset-level risk scores for urban infrastructure, and how can this be made accessible through a user-friendly interface?
Methodology
Google Earth Engine API integration for satellite imagery retrieval and processing. Geospatial machine learning models built with Scikit-learn. Risk scoring pipeline applied to transformer and utility asset datasets. Results rendered as interactive choropleth maps.
Analysis
Backend developed in FastAPI with geemap and Earth Engine Python API. Frontend built in Next.js 16 with React-Leaflet for interactive risk map rendering. Jupyter notebook documents the full GeoAI analysis pipeline.
Insights
Transformer assets in low-lying, high-density zones show elevated composite risk scores. The GeoAI pipeline demonstrates how Earth Engine-derived indices (NDVI, NDWI, built-up density) can augment ground-level asset data for risk classification.
Outcome
Deployed interactive risk map (transformer_risk_map.html) and full-stack web application. Reusable GeoAI pipeline documented in GeoAI_Equicity.ipynb for extension to other asset classes and cities.