iAmPhyton/Beyond-the-2.4-Mapping-the-Hidden-Paths-of-Nigerian-Finance

Analyzed microdata from the EFInA Access to Financial Services survey to investigate structural barriers in Nigeria's financial ecosystem. Barriers to Entry: A Livelihood Analysis of Nigerian Finance 🇳🇬 Project Overview - Recent reports indicate that only 2.4% of Nigerians earn above ₦200k per month. But income is only half the story. The bigger question is access: Who gets into the formal financial system, and who gets left behind? - This project analyzes the EFInA Access to Financial Services in Nigeria dataset to visualize the barriers to entry across livelihoods, geography, and education. Key Insights & Visualizations: 1. The "Informal Leak" (Sankey Diagram) Question: Do business owners bank formally? Finding: Contrary to expectation, a large segment of "Business Owners" flows directly into Informal Savings (Ajo/Esusu) rather than formal bank savings. This suggests that the formal banking sector is failing to capture the liquidity of Nigeria's informal economy. 2. The Shape of Inequality (Violin Plot) Question: What does wealth distribution actually look like? Finding: By plotting Wealth Scores against Livelihoods, we see that "Farming" has a heavy bottom density (mass poverty), while "Formal Labor" shows a "guitar shape" indicating a healthier middle-class distribution. 3. The Education Equalizer (Heatmap) Question: Is location destiny? Finding: Data shows that Education trumps Geography. A university graduate in the North East (typically a lower-access region) has higher banking penetration than an uneducated resident in the South West. Tools & Libraries: - Python: Core analysis. - Pandas: Data cleaning and SPSS (.sav) file handling. - Plotly: Interactive flow diagrams (Sankey). - Seaborn/Matplotlib: Statistical visualizations (Violins, Heatmaps). - Pyreadstat: For parsing complex survey metadata. Technical Challenges: One of the major challenges in this analysis was data granularity. Initially, I intended to build a "Financial Demographic Pyramid," but I discovered that the `Age` variable in the public dataset was pre-binned into broad categories (15-17 vs 18+), making granular age analysi …