Sally-52/Tanzania-Water-Wells-Project

# Tanzania Water Wells — Predictive Analysis > Predicting the functional status of water wells across Tanzania to support data-driven infrastructure decisions. --- ## Project Overview Access to clean water is a fundamental human right — yet thousands of water points across Tanzania are non-functional or in need of repair. This project uses real-world water well data to build a machine learning classification model that predicts whether a given well is **functional**, **non-functional**, or **in need of repair**. The goal: help governments and NGOs prioritize maintenance efforts and allocate resources more effectively. - ## Objectives - Clean and preprocess a large, messy real-world dataset - Perform exploratory data analysis (EDA) to uncover patterns in well functionality - Build and evaluate a classification model to predict well status - Visualize key findings to communicate insights clearly ## Dataset - **Source:** DrivenData — Pump it Up: Data Mining the Water Table - **Size:** 59,400 water point records across Tanzania - **Target Variable:** `status_group` — Functional / Non-Functional / Functional Needs Repair - **Features include:** GPS coordinates, installer, funder, water quality, quantity, pump type, construction year, and more ## Tools & Technologies | Tool | Purpose | |------|---------| | Python | Core programming language | | Pandas & NumPy | Data cleaning and manipulation | | Matplotlib & Seaborn | Data visualization | | Scikit-learn | Machine learning model building & evaluation | | Jupyter Notebook | Interactive development environment | --- ## Project Workflow ### 1. Data Cleaning - Handled missing values and zero-filled coordinates - Encoded categorical variables - Removed duplicate and irrelevant columns ### 2. Exploratory Data Analysis - Mapped well locations and functionality by region - Analyzed relationships between water quantity, quality, and pump status - Identified key features correlated with well failure ### 3. Mode …