Skip to content

Kuxha/disasterscout

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

23 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

DisasterScout

DisasterScout is a real‑time crisis‑intelligence tool that turns live news, geocoding, and LLM reasoning into an interactive disaster map and chat‑based assistant.

It helps users quickly understand what is happening during events like floods, storms, wildfires, and infrastructure emergencies — showing SOS zones, shelters, and general info reports directly on a map.

How It Works

The system automatically:

  • Queries Tavily for disaster‑related news.

  • Filters & Cleans results using OpenAI + custom NLP logic to classify incidents.

  • Geocodes extracted affected locations.

  • Deduplicates and stores data in MongoDB Atlas.

  • Displays incidents on a fast, interactive Leaflet map.

  • Responds to natural‑language messages (e.g., “Flood in Brooklyn, NY”) with summaries, guidance, and a link to the map.

Tech Stack

  • Backend: FastAPI

  • Database: MongoDB Atlas

  • Search & Intelligence: Tavily Search API, OpenAI (classification, filtering, LLM reasoning)

  • Frontend: LeafletJS map, Simple chat UI (no framework)

  • Data Pipeline: Custom ingestion pipeline with semantic/geo deduplication

How to Run

1. Install dependencies

Bash

pip install -r requirements.txt

2. Set environment variablesEnsure the following variables are set in your environment or .env file:

  • OPENAI_API_KEY

  • TAVILY_API_KEY

  • MONGO_URI

  • MONGO_DB_NAME

3. Start server

Bash uvicorn api_server.main:app --reload

4. Open the map UINavigate to: http://localhost:8000/map

Example Queries

You can ask the assistant queries such as:

"Flood in Brooklyn, NY"

"Flood in Bay Ridge, Brooklyn, NY"

"Flood in Qui Nhon, Vietnam"

The assistant returns a short situation report, actionable guidance, and a button to open the map centered on that region.

About

Crisis response platform for low-bandwidth environments. Geo-spatial implementation using MongoDB.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

No releases published

Packages

 
 
 

Contributors