DisasterIQ

Real-time disaster & disease outbreak risk intelligence

DisasterIQ predicts post-disaster disease outbreak risk for any location on Earth. Click anywhere on the map — it fetches live weather, USGS flood gauge readings, and historical precipitation data, then runs a SEIR epidemiological model combined with XGBoost classifiers to predict outbreak risk for 5 diseases in real time.

---

How it works

Click on map
    → Fetch live data (OWM + Open-Meteo + USGS)
    → Engineer 15 epidemiological features
    → Run SEIR simulation per disease
    → XGBoost classifies risk level (LOW → CRITICAL)
    → Return R₀, peak infected, total infected

Data sources

Source	What we use
OpenWeatherMap	Live temperature, humidity, rainfall, weather conditions
Open-Meteo	Historical daily precipitation (5-day lookback, free, no key)
USGS Water Services	146+ active river gauge readings for real flood detection

ML pipeline

• 4,090 training samples generated via SEIR simulation across 11 global locations
• 5 XGBoost classifiers (one per disease), each predicting LOW / MODERATE / HIGH / CRITICAL
• 15 engineered features including flood risk index, humidity-temperature interaction, days since heavy rain
• SEIR model (Susceptible → Exposed → Infected → Recovered) for epidemiological simulation with disease-specific R₀ parameters

Model performance

Disease	Accuracy	AUC
Cholera	97.7%	0.998
Dengue	96.9%	0.999
Malaria	96.9%	0.998
Leptospirosis	96.0%	0.997
Salmonella	97.8%	0.998

Scenario validation (15/15 correct)

Validated against documented historical outbreaks:

Scenario	Leptospirosis	Cholera	Malaria
Hurricane Harvey (Houston, 2017)	CRITICAL	HIGH	HIGH
Chennai Floods (India, 2015)	HIGH	HIGH	CRITICAL
Normal baseline	LOW	LOW	HIGH

---

Stack

Frontend

• Next.js 14 (App Router)
• Leaflet + React-Leaflet (map)
• Recharts (SEIR curves)
• Tailwind CSS

Backend

• FastAPI
• XGBoost
• NumPy / Pandas / SciPy
• Loguru

APIs

• OpenWeatherMap (current weather + 5-day forecast)
• Open-Meteo (historical precipitation archive)
• USGS Water Services (live flood gauges)
• Twilio (SMS alerts)

Deployment

• Frontend: Vercel
• Backend: Railway

---

Project structure

DisasterIQ/
├── backend/
│   ├── main.py           # FastAPI app — /predict, /alerts/subscribe
│   └── __init__.py
├── frontend/
│   ├── app/
│   │   ├── page.tsx      # Main map + panel layout
│   │   ├── layout.tsx    # Metadata, fonts
│   │   └── globals.css   # Design tokens
│   ├── components/
│   │   ├── Map.tsx       # Leaflet map with risk hotspot markers
│   │   ├── RiskPanel.tsx # Sidebar with metrics + disease cards
│   │   └── AlertModal.tsx# SMS alert subscription
│   └── types.ts          # Shared types + risk color helpers
├── ml/
│   ├── data_ingestion.py # OWM + Open-Meteo + USGS fetchers
│   ├── features.py       # Feature engineering (15 features)
│   ├── seir.py           # SEIR epidemiological model
│   ├── train.py          # XGBoost training pipeline
│   ├── evaluate.py       # Model evaluation + bootstrap CI
│   └── models/           # Trained XGBoost model weights
├── requirements.txt
└── railway.json

---

Running locally

Prerequisites: Python 3.11+, Node.js 18+

# Clone
git clone https://github.com/DevanshMalhotra17/Pathogen
cd Pathogen

# Backend
pip install -r requirements.txt
cp .env.example .env   # add your API keys
uvicorn backend.main:app --reload

# Frontend (new terminal)
cd frontend
npm install
npm run dev

Open http://localhost:3000

Environment variables

OPENWEATHERMAP_API_KEY=your_key   # openweathermap.org
TWILIO_ACCOUNT_SID=               # optional — for SMS alerts
TWILIO_AUTH_TOKEN=                # optional
TWILIO_FROM_NUMBER=               # optional

---

API

Interactive docs at /docs on the running backend.

GET  /health                        # Status + loaded models
POST /predict                       # { lat, lon } → risk predictions
GET  /predict/scenario/{name}       # houston_harvey | chennai_floods | normal_baseline
GET  /scenarios                     # List available scenarios
POST /alerts/subscribe              # Subscribe phone to SMS alerts

Example request:

curl -X POST https://pathogen-production-354d.up.railway.app/predict \
  -H "Content-Type: application/json" \
  -d '{"lat": 29.7604, "lon": -95.3698}'

Example response (Houston, TX):

{
  "lat": 29.7604,
  "lon": -95.3698,
  "usgs_flood_detected": true,
  "days_since_heavy_rain": 3,
  "risks": [
    {
      "disease": "leptospirosis",
      "risk_level": "CRITICAL",
      "confidence": 99.3,
      "r0": 4.72,
      "peak_infected": 117840,
      "total_infected": 136032
    }
  ]
}

---

Retrain the models

cd ml
python train.py      # generates ml/models/xgb_*.json
python evaluate.py   # generates ml/models/eval_results.json

---

Built for WeatherWise Hack 2026