DATA PATHWAYS TO HEALTHY CITIES AND HUMAN ASSESSMENT
BY: STARSHIP HACKERS DATE: OCTOBER 2025
Project Summary (English, 50–200 words) EarthChange is a two-level, Mexico-wide planning tool (Flask + Folium/Leaflet + Chart.js) that helps urban planners turn NASA Earth observations into concrete city actions. It ingests NASA FIRMS (active fires/smoke) and MODIS NDVI (vegetation cover; AppEEARS-ready, with a geographic fallback), alongside OpenWeather (heat/wind), WAQI/OpenAQ (PM2.5, NO₂, O₃), OSM green-space counts, and census/WorldPop density. For each state→city/municipality it computes a transparent Health & Resilience Score (≈30% air quality; 25% green/NDVI; 20% climate—temperature/humidity; 15% urban form—density/noise; 10% healthcare access) and surfaces risks such as wildfire proximity and low greenness. Planners, parks and health departments, and civil-protection teams can 1-. locate communities needing greener streets or shade, 2-. prioritize clinics and cooling resources during heat waves, 3-. monitor smoke and air-quality episodes, and 4-. track progress over time via dashboards and CSV/HTML exports. An AI module (Gemini) turns indicators into 12-month forecasts and actionable recommendations. The result is a reproducible, scalable pathway from NASA data to people- and planet-conscious urban growth.
EarthChange is an interactive planning tool that converts NASA Earth observations and open data into city actions for Mexico. It works as a two-level map from state to city or municipality. In near real time it queries NASA FIRMS for active fires and smoke and MODIS NDVI via Earthdata/AppEEARS for vegetation cover, combines this with OpenWeatherMap for temperature, humidity, and wind, OpenAQ and the WAQI service for air quality, and the OpenStreetMap Overpass API for parks and urban features. A Python pipeline without any database cleans and normalizes each stream on the fly, computes an explainable Health & Resilience Score on a 0–100 scale with visible weights for air, greenness, climate stress, urban form, and healthcare access, and renders results on a responsive Leaflet/Folium map with drill-downs, charts, CSV and HTML exports, plus a scenario mode to test interventions such as adding tree canopy, opening clinics, or applying traffic restrictions.
The immediate benefit is a single place to spot where vulnerabilities concentrate during heat waves, wildfire smoke events, or periods of poor air quality, and to locate neighborhoods lacking shade, parks, or health services. The intended impact is more equitable, climate-resilient growth: planners can prioritize green corridors, cooling centers, and clinic placement; parks and health departments can coordinate targeted actions; civil protection can monitor evolving risks; and leaders can track progress over time with transparent indicators.
The stack uses Python 3 and Flask for the backend, GeoPandas, Rasterio, and GDAL for geospatial processing, Folium/Leaflet and Plotly for visualization, environment variables for key management, and a lightweight task scheduler for periodic updates. It runs reproducibly with Docker, and needs only a standard laptop or a small cloud VM—no specialized hardware and no database.
Creatively, EarthChange bridges satellite science and municipal workflows through an explainable index and “why” cards that justify each recommendation, plus an AI module that generates 12-month forecasts and plain-language guidance. The team focused on data quality and recency, spatial bias and uncertainty, accessibility for low-bandwidth users, governance and privacy, scalability to additional cities, and meaningful involvement of city leaders and residents.