🐕‍🦺 Watch Dogs — Real-Time AI Emergency Intelligence

"In emergencies, information is life. Watch Dogs turns chaos into clarity."

💡 Inspiration

When disaster strikes, every second counts. But police officers, firefighters, and emergency responders face a critical problem: social media explodes with hundreds of unverified posts, making it nearly impossible to separate real emergencies from noise.

We witnessed this firsthand during recent California wildfires—officers were manually scrolling through X trying to piece together what was actually happening while lives hung in the balance. We asked ourselves:

What if AI could do that work instantly?

Watch Dogs was born from a simple truth: First responders need verified intelligence, not social media chaos. We built a system that turns X's real-time data stream into actionable emergency intelligence using Grok AI, giving police officers the clarity they need to save lives.

🚨 What It Does

Watch Dogs is an AI-powered emergency intelligence platform that monitors X in real time, verifies incidents using Grok, and delivers clean, actionable alerts to first responders.

🔧 Core Features

🔴 Real-Time Incident Detection

Streams location-tagged posts, trending emergency keywords, and verified accounts on X
Monitors police scanners, government alerts, and news feeds simultaneously
Detects fires, shootings, floods, and other emergencies as they unfold

🤖 Grok AI Verification Engine

Clusters related posts about the same incident
Cross-references X data with FEMA feeds, Cal Fire, LAPD, and news APIs
Generates confidence scores
- Example: "94% verified based on 47 X posts, 2 news reports, 1 gov alert"
Produces concise summaries in seconds

🗺️ Interactive Map Dashboard

Live incident map with color-coded severity markers
Pulsing alerts for active situations
Click markers to view full incident details
Location-based filtering for jurisdictions

📱 Social Media-Style Feed

Clean, scrollable incident cards with video previews
LIVE indicators for active situations
Source breakdown (X posts, news, government alerts)
One-tap access to full X threads and raw data

🎯 Severity Classification

CRITICAL: Active threats requiring immediate response
HIGH: Developing situations needing urgent attention
MEDIUM: Monitored events with moderate risk
LOW: Situational awareness updates

👮 For Police Officers

Stop doomscrolling through hundreds of unverified posts
Get verified summaries in under 30 seconds
See confidence scores so you know what to trust
Access raw sources to verify details yourself
Share intel instantly with dispatch and other units

⚙️ How We Built It

🧩 Technology Stack

Frontend (React)

Responsive web dashboard with dual view modes (feed + map)
Real-time UI updates using React hooks
Modular components for incident cards, map markers, and video players
Tailwind CSS for modern, mobile-friendly design

X API Integration

Streaming API for location-tagged emergency posts
Filtered streams for verified accounts
- @CalFire_NEWS, @LAPDHQ, @NWSLosAngeles
California geo-bounding box for regional monitoring
Rate limit handling with intelligent batching
Automated posting of verified alerts back to X

Grok AI Processing

Clustering algorithm: Groups related posts
Cross-reference engine: Validates against
- News APIs (Reuters, AP, local news)
- Government feeds (FEMA, National Weather Service)
- Official emergency accounts (Cal Fire, LAPD)
Confidence scoring weights:
- Government alerts = 100%
- Verified news = 80%
- Verified individuals = 50%
- Unverified posts = 20%
Summary generation: First-responder-optimized natural language