🇪🇹 Forecasting Financial Inclusion in Ethiopia
Interim Submission – Tasks 1 & 2
This project builds a data-driven system to forecast Ethiopia’s financial inclusion using time series data, national events, and policy interventions.
It follows the Global Findex framework and models two core pillars:
ACCESS — Account Ownership
USAGE — Digital Payment Adoption
đź“‚ Project Structure
ethiopia-fi-forecast/
├── data/
│ ├── raw/ # Starter dataset
│ ├── processed/ # Enriched + cleaned data
│
├── notebooks/ # Analysis notebooks
│ ├── 01_data_exploration.ipynb
│ └── 02_eda_analysis.ipynb
│
├── reports/
│ ├── figures/ # Charts from EDA
│ └── data_enrichment_log.md
│
├── dashboard/ # (Future task)
├── models/ # (Future task)
├── src/
├── requirements.txt
├── README.md
└── .gitignore
đź§ Unified Data Design
This dataset follows a neutral schema:
Events are NOT assigned to pillars.
Their effects are captured using impact_link records.
record_type category pillar Meaning
observation (empty) YES Measured indicator
target (empty) YES Policy goal
event YES (empty) Neutral event
impact_link (empty) YES Event → indicator effect
This avoids bias and allows one event to affect many pillars.
🧪 Task 1 – Data Exploration & Enrichment
âś” Completed
Loaded unified dataset
Reviewed schema & reference codes
Analyzed:
record types
pillars
sources
confidence levels
time coverage
Added new indicators & events
Documented all changes in: