eboekenh/pumprediction

Interactive Streamlit ML app for Tanzania water pump prediction — original prototype (see PumpPredictor for refactored version) # 💧 Pump Prediction — Interactive Streamlit ML Application > **Note:** This is the original interactive prototype built on Replit. For the refactored, modular ML pipeline version, see PumpPredictor. Interactive Streamlit application for predicting water pump functionality in Tanzania using machine learning, featuring an embedded EDA dashboard, feature engineering controls, and model training UI. ## Features - **Interactive EDA Dashboard** — Visual exploration of pump data with charts and statistics - **Feature Engineering UI** — Configure preprocessing steps through the Streamlit interface - **Model Training** — Train and evaluate ML models with real-time feedback - **Flask API** — REST API endpoint for model predictions - **Sample Data** — Included sample dataset for immediate testing ## Documentation This repository includes extensive documentation: | Document | Description | |----------|-------------| | `COMPLETE_CODE_DOCUMENTATION.md` | Full technical documentation | | `EDA_ANALYSIS_SUMMARY.md` | Exploratory data analysis findings | | `FEATURE_ENGINEERING_ACTIONS.md` | Feature engineering methodology | | `MODEL_SELECTION_AND_TUNING_ANALYSIS.md` | Model comparison and tuning results | | `VARIABLES_DETAILED_ANALYSIS.md` | Detailed variable analysis | | `BEGINNER_DATA_SCIENCE_GUIDE.md` | Learning guide for beginners | ## Related Repository - **PumpPredictor** — Clean, modular ML pipeline version with tests and CI-ready structure ## License MIT