Josemuturi/AI-Disease-Outbreak-Detection-System

A machine learning-powered surveillance system designed to predict and visualize infectious disease outbreaks (Malaria & Cholera) in Kenya using LSTM neural networks and geospatial mapping. # AI-Driven Disease Outbreak Surveillance System 🇰🇪 ## 📌 Project Overview A full-stack predictive analytics platform designed to monitor and forecast disease outbreaks across Kenyan counties. By leveraging historical health data and environmental factors, the system provides real-time risk assessments to aid public health decision-making. ### Key Features: * **AI Inference Engine:** Uses a pre-trained **LSTM (Long Short-Term Memory)** neural network for time-series forecasting. * **Interactive Dashboard:** A modern Next.js interface for visualizing risk levels (e.g., Nairobi: 95%). * **Geospatial Metadata:** Integrated GIS coordinate systems (EPSG:4326) for mapping county boundaries. * **RESTful API:** A high-performance FastAPI backend with automated OpenAPI (Swagger) documentation. --- ## 🏗️ System Architecture The system follows a **Decoupled Architecture**, separating the heavy computational logic (AI) from the user interface. ### 1. Frontend (The Face) * **Framework:** Next.js 14 (App Router) * **Styling:** Tailwind CSS * **State Management:** React Hooks (useEffect, useState) for handling asynchronous API calls and loading spinners. ### 2. Backend (The Brain) * **Framework:** FastAPI (Python) * **Server:** Uvicorn (ASGI) * **AI Logic:** TensorFlow/Keras for loading and running the `.h5` model. * **Data Handling:** Pandas for processing CSV-based health surveillance records. --- ## 🚀 Getting Started ### Backend Setup 1. Navigate to `/backend` 2. Install dependencies: `pip install -r requirements.txt` 3. Run the server: `python -m uvicorn main:app --reload` 4. Access API Docs: `127.0.0.1` ### Frontend Setup 1. Navigate to `/frontend` 2. Install modules: `npm install` 3. Run development server: `npm run dev` 4. Access Dashboard: `localhost` --- ## 📊 Data & Model Methodology * **Model Architecture:** Recurrent Neural Network (RNN) using LSTM layers. * **Training Data:** Historical infection rates and climate variables sto …