Logo Lanfrica

game-ale/ethiopia-fi-forecast

Domain:

socioeconomicdigital infrastructure

Record type:

project
Creator:
gam
Host:
# πŸ‡ͺπŸ‡Ή Ethiopia Financial Inclusion Forecasting System > "A comprehensive forecasting system tracking Ethiopia's digital financial transformation (2011–2027) using time-series analysis and event-impact modeling." --- ## πŸ“Š Overview Ethiopia is undergoing a rapid digital financial transformation. While mobile money registration has surged (Telebirr >54M, M-Pesa >10M), Global Findex demand-side data shows a complex picture of account ownership growth. This project builds a predictive system for the **Consortium of Stakeholders** to understand the drivers behind these trends and project inclusion outcomes for **2025–2027**. ### 🎯 Key Objectives 1. **Enrich** legacy datasets with leading infrastructure and enabler metrics (Smartphones, 4G, Digital ID). 2. **Analyze** the "Ownership Paradox" and the P2P-to-ATM crossover. 3. **Model** the lagging impacts of policy shifts (FX Liberalization, Open Banking) and product launches. 4. **Forecast** Access and Usage indicators with high/medium/low scenarios. --- ## πŸ— System Architecture ```mermaid graph TD A[Raw Data: Findex, Operator Reports, NBE] --> B[Data Enrichment Pipeline] B --> C[Unified Data Schema] C --> D[Exploratory Data Analysis] D --> E[Event Impact Modeling] E --> F[Forecasting Engine] F --> G[Interactive Streamlit Dashboard] subgraph "Unified Schema" C1[Observations] C2[Events] C3[Impact Links] C4[Targets] end ``` --- ## πŸ“‚ Project Structure ```text ethiopia-fi-forecast/ β”œβ”€β”€ data/ β”‚ β”œβ”€β”€ raw/ # Enriched unified data (CSV) β”‚ └── processed/ # Feature-engineered & Forecast data β”œβ”€β”€ notebooks/ β”‚ β”œβ”€β”€ EDA.ipynb # Interactive deep-dive into inclusion patterns β”‚ β”œβ”€β”€ impact_modeling.ipynb # Validation of event impacts β”‚ └── forecasting.ipynb # Scenario visualization β”œβ”€β”€ src/ β”‚ β”œβ”€β”€ init_data.py # Excel-to-Unified-CSV conversion β”‚ β”œβ”€β”€ enrich_data.py # Pipeline for adding new proxy observations β”‚ β”œβ”€β”€ impact_modeling.py # Logic for qu …