-
-
Manage, track, and update disaster incidents through a centralized interface.
-
AI-generated summaries and recommendations for faster emergency response decisions.
-
Interactive map displaying disaster locations and real-time incident information.
-
Real-time dashboard for monitoring disaster incidents and response priorities.
-
AI analyzes disaster reports and generates structured emergency insights instantly.
🌍 GroundTruth – AI-Powered Disaster Intelligence Platform
About the Project
Natural disasters often create a flood of information from citizens, emergency responders, and local authorities. Most of these reports are unstructured, making it difficult for response teams to quickly identify critical incidents, allocate resources, and understand the overall situation. Delays in processing this information can slow emergency response when every minute matters.
GroundTruth was built to address this challenge by transforming unstructured disaster reports into structured, actionable intelligence using artificial intelligence. Instead of manually reviewing lengthy reports, responders receive a concise summary with the disaster type, location, priority level, estimated impact, resource requirements, and AI-generated recommendations.
Inspiration
The idea was inspired by the growing frequency of floods, earthquakes, landslides, and industrial accidents, where emergency teams often receive hundreds of reports within a short period. We wanted to build a platform that could assist disaster management authorities by reducing information overload and helping them focus on the incidents that require immediate attention.
How We Built It
GroundTruth consists of a modern full-stack architecture:
- Frontend: React, TypeScript, Vite, Tailwind CSS, React Router, React Leaflet, OpenStreetMap
- Backend: FastAPI, SQLAlchemy, SQLite
- AI: Groq API with the GPT OSS model for intelligent disaster report analysis
The workflow is straightforward:
- A disaster report is submitted through the web interface.
- The AI analyzes the report and extracts structured information such as:
- Disaster type
- Location
- Priority
- Number of affected people
- Immediate resource requirements
- Confidence score
- Response recommendations
- The processed incident is stored in the database.
- The dashboard updates in real time with incident statistics.
- Incidents are displayed on an interactive map for better situational awareness.
- Emergency teams can track incident status from Active to In Progress and finally Resolved.
- The AI Command Center provides an overall situation summary and highlights the most critical incidents.
Challenges We Faced
One of the biggest challenges was designing a reliable AI workflow that consistently converted free-form disaster descriptions into structured data suitable for emergency response. We also worked on integrating multiple components—including AI analysis, database persistence, dashboard analytics, and geospatial visualization—into a seamless user experience.
Another challenge was ensuring that the application remained responsive while keeping the interface intuitive for users who may need to make quick decisions during emergency situations.
What We Learned
Building GroundTruth provided valuable experience in integrating large language models into real-world applications. We gained hands-on knowledge of full-stack development with React and FastAPI, REST API design, database management using SQLAlchemy, prompt engineering, and interactive geospatial visualization using Leaflet.
More importantly, we learned how AI can be applied beyond conversational use cases to solve practical problems by converting complex, unstructured information into meaningful decision support.
Future Scope
GroundTruth has the potential to evolve into a comprehensive disaster management platform by incorporating:
- Live weather and flood forecasting APIs
- Satellite imagery integration
- Drone-based damage assessment
- Multi-language disaster report analysis
- SMS and emergency notification systems
- GPS-based incident reporting
- Resource allocation optimization
- Predictive disaster risk analytics
Our vision is to make GroundTruth a reliable decision-support platform that empowers emergency responders with timely, accurate, and actionable intelligence during disasters.
Built With
- ai
- axios
- disastermanagement
- fastapi
- gpt-oss
- groq
- machine-learning
- openstreetmap
- python
- react
- reactleaflet
- reactrouter
- restapi
- sqlalchemy
- sqlite
- tailwindcss
- typescript
- vite
Log in or sign up for Devpost to join the conversation.