christaingabire/public-health-etl

ETL pipeline project analyzing public health indicators in Sub-Saharan Africa from 1960 to 2023 # Public Health Data ETL Pipeline (Sub-Saharan Africa) ## Overview This project explores how public health outcomes evolved across Sub-Saharan Africa over six decades. It automates the process of transforming raw World Bank data into an analytics-ready database, enabling longitudinal comparison of life expectancy, immunization, and maternal health indicators. It reflects my early work on building data infrastructure for public systems; A foundation that later evolved into large-scale social protection analytics at RSSB. ## Introduction This project builds a full ETL (Extract, Transform, Load) pipeline using public health data from the World Bank's Health Nutrition and Population Statistics database. It focuses on analyzing three key health indicators across four Sub-Saharan African countries from 1960 to 2023: - Life Expectancy at Birth - Immunization Coverage (DPT3 vaccine) - Maternal Mortality Ratio This project reflects my interest in using data engineering to analyze public infrastructure, development, and social systems. I wanted to practice building a complete data pipeline- from messy raw files to a structured, clean database- while exploring how health outcomes have evolved over time across different regions of Africa. ## Dataset Information - Source: World Bank - Health Nutrition and Population Statistics - Format: Excel file (.xlsx) and JSON metadata - Time Period: 1960–2023 - Countries Focused: Kenya, Ghana, South Africa, Ethiopia --- ## Technologies Used - Python3 - Pandas - SQLite3 - JSON - Excel file handling ("openpyxl") ## How the Pipeline Works 1. **Extract**: - Read the Excel and JSON files containing public health data. 2. **Transform**: - Filtered data for the selected specific countries and indicators. - Reshaped the data into tidy format data (years as rows). - Dropped missing values and cleaned the dataset for effective analysis. 3. **Load**: - Saved the cleaned data to a `.csv` file. - Loaded the final data into an SQLite databa …