BalogunEmmanuelKehinde/Africa-remittance-analysis

# 🌍 Africa Remittance Cost Analysis An end-to-end data pipeline and investigative analytics project examining the true cost of sending money to, from, and within Africa — using real World Bank data spanning 2016 to 2025. --- ## 📌 The Finding That Started Everything > A Tanzanian bank charges **91%** to send money to Uganda. > Western Union charges **8%** on the exact same route. > Same corridor. Same quarter. 10x the price. This project was built to understand why — and whether Africa is on track to meet the UN SDG 10.c target of **3% remittance costs by 2030**. **Spoiler: The trend is going the wrong direction.** --- ## 📊 Dashboard Preview ### Page 1 — The Big Picture ### Page 4 — The Deep Dive > Built in Power BI, connected live to PostgreSQL. --- ## 🏗️ Project Architecture ``` World Bank Excel (47,000+ rows) ↓ clean_data.py ← Python + Pandas (Extract, Transform) ↓ PostgreSQL ← africa_remittances table (Load) ↓ main.py (FastAPI) ← REST API serving analytics endpoints ↓ Power BI Dashboard ← 4-page live visualization layer ``` --- ## 📁 Repository Structure ``` africa-remittance-analysis/ │ ├── clean_data.py # Data cleaning & ingestion pipeline ├── check_sheets.py # Data validation & column inspection ├── main.py # FastAPI application ├── queries.sql # Key analytical queries ├── africa_remittances_clean.csv # Cleaned Africa-filtered dataset └── README.md ``` --- ## ⚙️ Pipeline Breakdown ### 1. Ingestion & Cleaning (`clean_data.py`) - Loads the World Bank Remittance Prices Worldwide dataset (Excel) - Filters for all Africa-related corridors bidirectionally — rows where Africa is either the source or destination - Removes promotional noise (negative cost percentages) that would distort analysis - Classifies each transaction into one of three flow types: - `Intra-Africa` — both source and destination are African countries - `Outbound (Africa to …