NickMwai/project

tanzania water model # Water Pump Functionality Prediction A machine learning project that predicts the operational status of water pumps in rural Tanzania using metadata. The goal is to support proactive maintenance and ensure sustainable water access. ## Project Overview This project uses the Tanzania Water Pump dataset from DrivenData. It applies data preprocessing, feature engineering, and classification models to predict whether a pump is: - Functional - Functional but needs repair - Non-functional Achieved **81% accuracy** using a Random Forest Classifier. ## Tech Stack / Tools - Python (Pandas, scikit-learn, matplotlib, seaborn) - Jupyter Notebook - XGBoost / RandomForest - SMOTE (imbalanced-learn) ## How to Run the Project 1. Clone the repo: ```bash git clone github.com cd water-pump-predictor ``` 2. Install dependencies: ```bash pip install -r requirements.txt ``` 3. Run the notebook: ```bash jupyter notebook notebooks/Water_Pump_Model.ipynb ``` 4. (Optional) Predict on new data: ```python model.predict(new_data) ``` ## Project Structure ``` 📁 data/ └── Training_Set_Values.csv └── Training_Set_Labels.csv └── Test_Set_Values.csv 📁 notebooks/ └── Water_Pump_Model.ipynb 📄 README.md 📄 requirements.txt ``` ## 📈 Results - Accuracy: 81% - Confusion Matrix: - Functional: Precision 0.81, Recall 0.89 - Needs Repair: Precision 0.55, Recall 0.33 - Non-functional: Precision 0.84, Recall 0.78 Visualizations and model performance plots available in the notebook. ## 📢 Business Impact - Enables data-driven pump maintenance in rural communities - Reduces downtime and repair costs - Improves infrastructure planning and public service delivery ## 🔮 Future Work - Improve "Needs Repair" classification using better features - Deploy model as a REST API - Integrate with GIS dashboards