Logo Lanfrica

Wa1r1muKariuki/Vee-GBV-Reporting-Platform

Domain:

natural language processingpeace and security

Record type:

software
Creator:
Wa1
Host:
A secure, bilingual (English/Kiswahili) web platform for anonymous GBV reporting, featuring a trauma-informed AI chatbot and real-time incident mapping. Built with Next.js, FastAPI, and PostgreSQL. # Vee-GBV-Reporting-Platform Vee is a secure, trauma-informed digital platform designed to bridge the gap in Gender-Based Violence (GBV) reporting in Kenya. It provides survivors with a safe, anonymous channel to report incidents, access verified support services, and contribute to data-driven safety mapping without fear of stigma or retaliation. ## Key Features - **Hybrid AI Chatbot**: A bilingual (English & Kiswahili) conversational interface powered by Google Gemini LLM and Rasa NLU. It uses trauma-informed scripts to guide survivors through reporting while prioritizing empathy and safety. - **Secure & Anonymous**: Features AES-256 field-level encryption for sensitive narratives. No personal identifiers are stored, and all data is transmitted via HTTPS. - **Incident Mapping**: Visualizes verified, anonymized reports on an interactive Mapbox heatmap to help communities and NGOs identify high-risk zones. - **Smart Referrals**: Uses PostGIS geospatial querying to instantly connect survivors with the nearest verified police stations, hospitals, and legal aid centers. - **Admin Dashboard**: A secure portal for authorized personnel to verify incoming reports and view aggregated analytics. ## Tech Stack - **Frontend**: Next.js (React), Tailwind CSS, Mapbox GL JS - **Backend**: FastAPI (Python), Pydantic - **Database**: PostgreSQL with PostGIS extension - **AI/ML**: Google Gemini API, Rasa Open Source (NLU) - **Security**: Python Cryptography (Fernet/AES) ## System Architecture The system follows a Three-Layer Architecture: 1. **Presentation Layer**: Next.js handles the responsive, mobile-first UI. 2. **Application Layer**: FastAPI manages the business logic, encrypts data, and orchestrates the hybrid AI workflow (Intent Recognition → Response Generation). 3. **Data Layer**: PostgreSQL stores encrypted reports and geospatial resource data. ## Installation To set up the project locally: ### 1. Clone the Repository ```bash git clone github.com