diawaraousmane63-hash/FinScore

A prototype AI model for alternative credit scoring in West Africa using simulated data # FinScore A prototype AI model for alternative credit scoring in West Africa using simulated data ## The Problem Over 60% of small businesses in West Africa lack access to formal credit. Traditional banks cannot assess their risk because they lack conventional banking history. ## The Vision FinScore aims to leverage alternative data (mobile money transactions, telco data) to build an AI-powered creditworthiness score for the informal economy. ## About This Repository This is a **working prototype and learning sandbox**. It represents my hands-on journey to technical proficiency as I build towards solving this problem. The work here is foundational and evolves as I learn. ### Project Structure - `data/`: Contains simulated mobile money transaction data. - `notebooks/`: - `01_data_simulation.ipynb`: Data generation and exploratory analysis. - `02_feature_engineering.ipynb`: Creating meaningful features from raw data. - `03_model_prototype.ipynb`: A simple ML model proof-of-concept. ## Next Steps & Learning Goals - [ ] Experiment with more advanced models (e.g., Gradient Boosting). - [ ] Integrate more data sources (simulated telco top-up patterns). - [ ] Build a simple Streamlit demo app. ## Connect With Me linkedin.com