SmogSense City Intelligence

AI-Powered Air Quality Forecasting & Recess Safety System — Central Lahore

Lahore Central Station (US Consulate)
Hackathon Live Pitch Mode:
Tomorrow's School Activity Directive
XGBoost Decision Engine • 100 µg/m³ Threshold

Analyzing Air Quality...

Connecting to FastAPI backend at http://127.0.0.1:8000 to compute next-day PM2.5 forecast and safety advisory...
POLICY RULE: PM2.5 > 100 µg/m³ = Mandatory Indoor Activity
Forecast Target: Tomorrow
Current PM2.5
-- µg/m³
Real-time Sensor
Predicted Tomorrow
-- µg/m³
XGBoost Forecast
24h Trajectory
-- µg/m³
Trend Analysis
Custom Safety Threshold
100.0 µg/m³ cutoff
Project-Defined Safety Threshold (Informed by AQ Guidance)

PM2.5 Trajectory & Next-Day AI Forecast

Recent historical observations + XGBoost predicted point for tomorrow with safety boundary

Observed PM2.5
Tomorrow's AI Prediction
Safety Threshold (100 µg/m³)

Decision Protocol for Administrators

Designed specifically for Lahore school principals, sports coaches, and administrative heads. Morning decisions are delivered with high lead time so PE schedules and bus routes can adjust before school buses depart.

Decision Predicted PM2.5 Directive
Proceed ≤ 100 µg/m³ Normal outdoor recess, sports & PE
Move Indoors > 100 µg/m³ Relocate all physical activities indoors

City Intelligence Architecture

Trained on multi-year Lahore atmospheric telemetry combining OpenAQ sensor historical lags (1d, 3d, 7d rolling means and 7d volatility) with Open-Meteo meteorological variables (temperature, humidity, wind velocity, precipitation).

XGBoost Classifier + Regressor Bimodal Smog Distribution Aware FastAPI Microservice