TheKingOfLogic/IBM-Z-Datathon-Project---AmanziGuard

A predictive water quality monitoring system for South African catchments, developed for the IBM Z Datathon 2025. Uses AI models to analyze physicochemical and trace-metal data for the Luvuvhu and uMhlathuze regions. # AmanziGuard ## Screenshot --- ## Problem Water quality in South African catchments, particularly the Luvuvhu and uMhlathuze regions, is being degraded by chemical, nutrient, and heavy-metal contamination from industrial, agricultural, and natural sources. Monitoring these waters manually is inefficient and slow, creating a need for an automated system that can predict water quality accurately. --- ## Solution AmanziGuard is a predictive water quality monitoring system developed for the IBM Z Datathon 2025. It uses machine learning to classify water samples as **Safe**, **Moderate**, or **Contaminated** based on physicochemical and trace-metal data. Key features include: - Cleaning and preprocessing water quality data using Python and Pandas - Training a **Random Forest Classifier** to predict water quality classes - Visualizing model performance with confusion matrices and classification reports using Matplotlib and Seaborn - Providing a web interface using **Flask**, allowing users to input water sample data and receive predictions - Storing and loading trained models with Joblib for easy deployment --- ## Tools - Python - Pandas - Scikit-learn - Matplotlib and Seaborn - Flask - Joblib - HTML/CSS