Water Quality Prediction Using Machine Learning

Project Overview

This project aims to build a robust machine learning classifier that predicts the safety of water samples based on the concentration of various chemical and biological contaminants. The classification model helps determine whether the water is safe (is_safe = 1) or unsafe (is_safe = 0) using features like heavy metals, nitrates, bacteria, viruses, and more.

---

Technical Highlights

• Dataset: Contains 7,999 water samples with 21 attributes including:

  • aluminium, ammonia, arsenic, lead, bacteria, viruses, and a binary target label is_safe.

• Data Cleaning:

  • Removed invalid entries (e.g., rows with #NUM!)
  • Handled non-numeric values using type conversion and dropped NaNs

• Visualization Techniques:

  • Histograms, boxplots, pairplots, heatmaps, violin plots, and scatterplots to analyze distribution, relationships, and correlations between contaminants

• Modeling:

  • Implemented and compared multiple classifiers: Logistic Regression, Random Forest, and XGBoost
  • Achieved best performance with XGBoost, followed by hyperparameter tuning using RandomizedSearchCV

• Performance:

  • XGBoost model achieved a cross-validated accuracy of 96.9%
  • Final evaluation using confusion matrix for visual performance analysis

---

Purpose and Applications

• Assess real-world water quality in environmental monitoring
• Aid governmental health agencies and municipal bodies in water safety checks
• Used in IoT-based water sensors for real-time water safety alerts
• Serves as a teaching aid for classification, data cleaning, and model tuning in machine learning

---

Installation

Clone the repository

git clone https://github.com/BhaveshBhakta/Water-Quality-Prediction-Using-XGBClassifier.git
cd Water-Quality-Prediction-Using-XGBClassifier

---

Collaboration

Contributions are welcome! If you’d like to improve model performance, add new visualizations, or integrate the project with a web interface.