Home / Projects / Chhaya: Climate-Aware Routing Engine
AI / Microclimate Routing

Chhaya: Climate-Aware Routing Engine

Global (6 Mega-cities) 2026

A 'Smart Google Maps for Walking' (adopted from the DETOUR project) that optimizes for human comfort and safety by calculating routes based on high-resolution microclimates, building shadows, wind tunnels, and personalized user personas.

ReactFastAPINetworkXSupabasePostGISMicroclimate
Chhaya Interactive App

Interactive map — scroll and click to explore.

Context

Traditional routing engines prioritize the shortest physical distance, often exposing pedestrians to extreme urban heat, blazing sun, and uncomfortable microclimates in mega-cities like Dubai, Delhi, and Barcelona.

Problem Statement

How can we integrate high-resolution environmental raster data (UTCI, solar radiation, wind speed) with vector street networks to mathematically calculate the most thermally comfortable and safe walking route for different user demographics?

Methodology

Adopted from the DETOUR project (developed for the Infrared Hackathon), this custom routing engine overlays OpenStreetMap street networks (stored in Supabase PostGIS) with 2D microclimate rasters retrieved from the Infrared.city AI oracle. Routes are scored using A* search based on dynamic persona weights (Athlete, Elderly, Tourist) that prioritize shade, nature, and cooling wind.

Analysis

Backend developed in Python (FastAPI, NetworkX, GeoPandas) running on a 16GB Hugging Face Space. Spatial database managed via Supabase (PostGIS) holding over 1 million pre-processed street edges. Frontend built in React + Vite with Mapbox GL JS for interactive heat-map rendering.

Insights

Routing based on microclimate drastically alters path selection. Incorporating building shadows and park locations often increases route distance slightly but significantly decreases the overall 'thermal cost', leading to vastly more comfortable walking experiences during peak heat hours.

Outcome

Deployed a full-stack, free-tier application serving 6 global mega-cities. The system successfully processes over 400,000 edges per city and instantly generates climate-optimized routes with visual heat-map overlays and Gemini-powered natural language POI injections.

Empowering Governance through
Spatial Intelligence

© 2025 Bharat Oraon. Urban Planning Portfolio.

Built with Love ❤️