Logo Lanfrica

Chege-jr0/PESA_Insight

Domain:

socioeconomic

Record type:

softwareproject
Creator:
Che
Host:
This is an AI_Powered platform that tracks M-pesa adoption, financial inclusion gaps and exclusion barriers across Kenya, with forecasting to 2030 and AI advisor for SME,s and policy makers. ## PesaInsight - Kenya Mobile Mney and Financial Inclusion An AI-powered intelligence platform that tracks M-pesa adoption, financial inclusion gaps and exclusion barriers across Kenya, with forecasting to 2030 and an AI adviser for SME,s and policy makers "Kenya leads the world in mobile money, Yet a minority of Kenyans remain completely financially excluded and in remote areas, the percentage is even higher. PesaInsights exists to make the contradiction visible, understable and actionable." ## Data Sources 1. 2025 FinAccess Household Survey 2. World Bank Global Findex Database 3. Safaricom Annual Reports 4. CBK Financial Sector Reports ## TechStack 1. Apache Airflow 3.0 => Automated weekly data updates 2. SQLite => Zero setup, portable and fast database. 3. Pandas => Cleaning, aggregation, feature engineering. 4. Numpy => Realistic county-level data generation. 5. Scikit-learn => Forecasting through linea regression for 2030 projections. 6. Streamlit => Interactive Dashboard in pure python. 7. Plotly Express => Interactive, professional visualisations charts 8. Ollama + TinyLlama => Free, private local AI to give predictions that worsk offline. 9. Python => Core language that ties everything together. ## Project Structure ```markdown pesa-insight/ │ ├── dags/ │ └── finance_pipeline.py # Airflow 3.0 DAG │ ├── data/ │ └── raw/ # Raw extracted CSV files │ ├── database/ │ └── finance.db # SQLite database (auto-created) │ ├── src/ │ ├── extract.py # Step 1 — Pull data from sources │ ├── transform.py # Step 2 — Clean & enrich │ ├── load.py # Step 3 — Write to SQLite │ ├── forecast.py # Step 4 — Generate 2030 projections │ └── ai_advisor.py # AI insights, SME & personal advisor │ ├── dashboard/ │ └── app.py # Streamlit dashboard │ ├── requirements.txt └── README.md ``` ## Key Metrics 1. Inclusion_rate => Percenta …