almanjenga/menstrual-health-chatbot

AI-powered web chatbot providing accurate, empathetic, and culturally sensitive menstrual health guidance for Kenyan youth, in English and Kiswahili. # Menstrual Health Chatbot (Kenyan Youth–Focused) A web-based, AI-powered menstrual health chatbot designed to provide **accurate, empathetic, and culturally sensitive guidance** to adolescents and young women in Kenya. The chatbot leverages **NLP, RAG (Retrieval-Augmented Generation), and sentiment analysis** to deliver reliable menstrual health information in both **English and Kiswahili**. --- ## 🌸 Project Overview Access to youth-friendly menstrual health education in Kenya is limited due to stigma, cultural taboos, and misinformation. This project develops a **React + Next.js web application** with a **Flask backend**, enabling users to: - Ask questions about menstrual health - Receive contextually accurate, culturally sensitive answers - Interact in **English or Kiswahili** - Benefit from emotionally aware responses The system combines a **fine-tuned Flan-T5 model** with a **vector-based RAG pipeline (FAISS)** to ensure precise and reliable information delivery. --- ## ⚡ Key Features - **Conversational Menstrual Health Support**: Guidance on cycles, hygiene, symptoms, and conditions. - **Cultural & Linguistic Sensitivity**: Supports **Kiswahili** and local terminology, making the chatbot relatable to Kenyan youth. - **Emotionally Intelligent Responses**: Uses sentiment analysis to adapt tone and provide empathetic support. - **RAG-Based Knowledge Retrieval**: Combines curated datasets with semantic search to provide **factually accurate answers**. - **Modern Web Interface**: Built with **React + Next.js** and styled with **Tailwind CSS** for a responsive, user-friendly experience. - **Firebase Authentication**: Secure login and personalized user experience. --- ## 🏗️ Project Structure ``` menstrual_health_chatbot/ ├── backend/ │ ├── model/ │ │ ├── config.json │ │ ├── generation_config.json │ │ ├── special_tokens_map.json │ │ ├── tokenizer.json │ │ └── tokenizer_config.json │ ├── TRANSLATION_README.md │ ├── app.py │ ├── build_swahili_index.py │ ├── …