Inspiration π
Every woman has faced that moment of hesitation before walking alone. The mental calculations, keys clutched between fingers, and a rushed pace as darkness fallsβthese aren't just inconveniences; they're violations of our fundamental right to move freely. As a woman who has encountered everything from subtle discomfort to outright harassment, I created Go Guardian because safety shouldn't be a privilege; it should be a guarantee π.
What It Does π
Go Guardian is an AI-powered safety companion that transforms how women navigate urban spaces:
Intelligent Route Analysis πΊοΈ
- Real-time safety scoring using Google's Gemini AI π
- Comprehensive analysis of street lighting and infrastructure π‘
- Identification of safe spaces along routes π‘οΈ
- Dynamic risk assessment based on time of day β°
- Weather impact analysis on route safety β
Community Safety Network π€
The community features include:
- Neighborhood safety groups ποΈ
- Real-time community alerts for construction, outages, and incidents π§
- Verified safety updates from local moderators π
- Emergency contact integration π
- Group-based safety monitoring π₯
Infrastructure Monitoring π
The safety monitoring includes:
- Real-time street light status tracking π
- Infrastructure coverage analysis ποΈ
- Area safety scoring π
- Working lights monitoring π¦
- Safe spaces density mapping π
How I Built It π οΈ
Technical Foundation
- Next.js 14 frontend with real-time safety visualization π±
- Python backend with Gemini AI integration π
- San Francisco city data integration for:
- Police incidents π
- Street lighting π‘
- Business locations πͺ
- Transit stops π
Safety Analysis System
The backend analyzes multiple factors:
- Time-based risk assessment (day/night analysis) π
- Infrastructure status (lighting coverage) π
- Historical incident patterns π
- Safe spaces density π¦
- Emergency resource proximity π
Challenges I Ran Into β οΈ
Real-time Analysis
- Balancing comprehensive safety analysis with response time β³
- Managing API rate limits with city data π
- Implementing reliable fallback systems π οΈ
- Coordinating multiple data sources π‘
Community Features
Currently a dummy page, with plans for features yet to be added.
Infrastructure Monitoring
- Building a database with SF data ποΈ
- Real-time light coverage analysis π
- Accurate safety score calculations β
- Handling data gaps and inconsistencies π
Accomplishments That Iβm Proud Of π
Technical Achievement
- Successfully integrated Gemini AI for contextual safety analysis π
- Brought in weather information β
- Attempted to incorporate voice and video features for wake-word recognition and emergency activation π€
- Created an intuitive community alert system π²
- Implemented real-time infrastructure monitoring π
Community Impact
- Developed neighborhood-specific safety groups π‘
- Created a verified alert system β
- Built infrastructure monitoring tools π§
- Fostered a supportive safety network π€
What I Learned π
Technical Insights
- Structured prompt engineering for consistent responses π
- Efficient data processing for real-time analysis β‘
- API rate limiting implementation π¦
- Infrastructure monitoring complexities ποΈ
Safety Analysis
- The importance of contextual safety data π‘οΈ
- The role of infrastructure in urban safety ποΈ
- The power of community-driven safety networks π€
- The complexity of real-time safety monitoring β°
- Time-sensitive route recommendations β
What's Next for Go Guardian π
- Beta testing with women π§ͺ
- Application scaling initiatives π
Voice Activation System
- Implementation of "Hi Shield!" wake word detection β
- Voice-activated emergency protocols π
- Hands-free safety commands:
- "Hi Shield! Call emergency contacts" π±
- "Hi Shield! Share my location" π
- Multi-language voice command support π
- Noise-resistant voice recognition π§
- Custom wake word training options π
Enhanced Gemini AI Integration π
- Advanced visual threat detection using Gemini Pro Vision π¬
- Real-time crowd behavior analysis π₯
- Suspicious activity recognition π¨
- Environmental hazard detection πͺοΈ
- Multi-modal safety assessment π¨
- Combined audio-visual analysis π₯
- Contextual environment understanding π
- Behavioral pattern recognition π
- Predictive safety analytics π
- Route risk forecasting πΊοΈ
- Incident probability modeling π
- Dynamic safety score adjustments βοΈ
- Natural language understanding improvements π£οΈ
- Context-aware emergency response π
- Emotional state analysis π
- Situation-specific guidance π
Go Guardian isn't just an app β it's a commitment to creating safer spaces for everyone. By combining AI technology with community engagement, we're working towards a future where safety isn't just a feature β it's a fundamental right. π
Built With
- fastapi
- git
- google-cloud-console
- google-gemini-pro
- google-maps
- next.js
- python
- shadcn
- tailwind
- typescript
Log in or sign up for Devpost to join the conversation.