acuestamd/geosentinel
GitHub: acuestamd/geosentinel
AI驱动的全球疾病监测网络，用OSINT实现实时疫情早期预警。
Stars: 2 | Forks: 1
# 🌐 GeoSentinel 2.0
**AI-Powered Global Disease Surveillance Network**
Real-time disease outbreak monitoring using open-source intelligence (OSINT). Replacing expensive, slow clinic-based surveillance with automated multi-source signal detection.
## 🛰️ Data Sources
| Source | Type | Coverage |
|--------|------|----------|
| 🏥 **WHO** | Disease Outbreak News API | Official alerts, global |
| 📰 **News** | Brave Search (CDC, Reuters, Al Jazeera...) | Breaking outbreaks |
| 𝕏 **Twitter/X** | Real-time social signals | Traveler reports, early detection |
| 💬 **Reddit** | Community reports | Travel health experiences |
| 📈 **Google Trends** | Search trend anomalies | Population-level signals |
## 🔬 Features
- **Real-time scanning** — automated every 30 minutes via GitHub Actions
- **Anomaly detection** — historical baseline comparison flags unusual spikes
- **Traveler signal detection** — NLP patterns identify "came back sick from..." reports
- **Flight risk modeling** — maps potential disease spread via air routes
- **Interactive map** — Leaflet-based with severity-coded markers
- **Multi-severity scoring** — 1-10 scale: critical, high, moderate, low
- **Source-filtered views** — filter by WHO, news, Twitter, Reddit, or Trends
## 🏗️ Architecture
```
GitHub Actions (cron 30 min)
→ scanner_v2.py (Python)
→ WHO API + Brave Search + Twitter + Google Trends
→ NLP disease detection + geocoding + anomaly scoring
→ signals.json
→ GitHub Pages (static deploy)
→ index.html (Leaflet map + dashboard)
```
**Zero infrastructure cost.** Runs entirely on GitHub Actions + Pages.
## 🚀 Deployment
This site auto-deploys via GitHub Pages. Every 30 minutes:
1. Scanner collects signals from 5 sources
2. Processes, deduplicates, scores, and geocodes
3. Commits `signals.json` to repo
4. GitHub Pages serves the updated dashboard
## 📊 Signal Processing Pipeline
1. **Collection** — parallel queries across 5 source APIs
2. **Disease Detection** — regex + NLP matching against 30+ disease patterns
3. **Geocoding** — 100+ country/city database with region classification
4. **Severity Scoring** — base disease severity + modifiers (deaths, outbreak scale, traveler)
5. **Anomaly Detection** — 2× historical baseline comparison
6. **Deduplication** — hash-based + location clustering
7. **Flight Risk** — IATA hub mapping for affected countries
## ⚕️ Background
The original [GeoSentinel](https://www.istm.org/geosentinel) is a WHO/CDC/ISTM clinic-based surveillance network — 70 clinics worldwide, slow reporting, limited coverage.
**GeoSentinel 2.0** replaces this with AI-powered OSINT: faster, broader, free, and available to everyone.
## 📜 License
MIT — Use freely. Attribution appreciated.
*Built with 🪁 by [Kite](https://github.com/acuestamd)*
## 🛰️ Data Sources
| Source | Type | Coverage |
|--------|------|----------|
| 🏥 **WHO** | Disease Outbreak News API | Official alerts, global |
| 📰 **News** | Brave Search (CDC, Reuters, Al Jazeera...) | Breaking outbreaks |
| 𝕏 **Twitter/X** | Real-time social signals | Traveler reports, early detection |
| 💬 **Reddit** | Community reports | Travel health experiences |
| 📈 **Google Trends** | Search trend anomalies | Population-level signals |
## 🔬 Features
- **Real-time scanning** — automated every 30 minutes via GitHub Actions
- **Anomaly detection** — historical baseline comparison flags unusual spikes
- **Traveler signal detection** — NLP patterns identify "came back sick from..." reports
- **Flight risk modeling** — maps potential disease spread via air routes
- **Interactive map** — Leaflet-based with severity-coded markers
- **Multi-severity scoring** — 1-10 scale: critical, high, moderate, low
- **Source-filtered views** — filter by WHO, news, Twitter, Reddit, or Trends
## 🏗️ Architecture
```
GitHub Actions (cron 30 min)
→ scanner_v2.py (Python)
→ WHO API + Brave Search + Twitter + Google Trends
→ NLP disease detection + geocoding + anomaly scoring
→ signals.json
→ GitHub Pages (static deploy)
→ index.html (Leaflet map + dashboard)
```
**Zero infrastructure cost.** Runs entirely on GitHub Actions + Pages.
## 🚀 Deployment
This site auto-deploys via GitHub Pages. Every 30 minutes:
1. Scanner collects signals from 5 sources
2. Processes, deduplicates, scores, and geocodes
3. Commits `signals.json` to repo
4. GitHub Pages serves the updated dashboard
## 📊 Signal Processing Pipeline
1. **Collection** — parallel queries across 5 source APIs
2. **Disease Detection** — regex + NLP matching against 30+ disease patterns
3. **Geocoding** — 100+ country/city database with region classification
4. **Severity Scoring** — base disease severity + modifiers (deaths, outbreak scale, traveler)
5. **Anomaly Detection** — 2× historical baseline comparison
6. **Deduplication** — hash-based + location clustering
7. **Flight Risk** — IATA hub mapping for affected countries
## ⚕️ Background
The original [GeoSentinel](https://www.istm.org/geosentinel) is a WHO/CDC/ISTM clinic-based surveillance network — 70 clinics worldwide, slow reporting, limited coverage.
**GeoSentinel 2.0** replaces this with AI-powered OSINT: faster, broader, free, and available to everyone.
## 📜 License
MIT — Use freely. Attribution appreciated.
*Built with 🪁 by [Kite](https://github.com/acuestamd)*标签：30分钟更新, AI驱动, Brave Search, ESC4, GitHub Actions, GitHub Pages, Google Trends, Leaflet地图, NLP, OSINT, Reddit, SEO, Twitter, WHO, 交互式地图, 低成本部署, 全球疾病监测, 公共卫生, 地理空间分析, 地理编码, 多源数据融合, 实时疫情监控, 异常检测, 新闻监测, 旅行者信号检测, 疾病预警, 自动笔记, 航班风险建模, 逆向工具, 静态网站