nyashaChiza/african-credit-scoring-challenge

Financial institutions need to predict loan defaults to mitigate risk and optimise lending decisions. In Africa’s rapidly growing financial markets, with diverse customer demographics and dynamic economic conditions, accurately assessing default risk is more important than ever. ### README: African Credit Scoring Challenge #### **Project Overview** This repository is designed for the African Credit Scoring Challenge, where the goal is to develop a machine learning model that predicts credit risk, improving access to financial services. #### **Repository Structure** - **`data/`** - `raw/`: Stores the original dataset files. - `processed/`: Contains processed iterations of the data for modeling. - **`models/`**: Stores trained model files. - **`pipelines/`**: Contains serialized preprocessing and modeling pipelines (`.pkl` files). - **`scripts/`**: Holds Python scripts for various tasks. - **`notebooks/`**: Jupyter notebooks for exploration, analysis, and experiments. - **`main.py`**: A FastAPI application for serving the trained model. #### **Problem Statement** Predict customer credit risk using anonymized financial data, enabling financial institutions to assess creditworthiness accurately. zindi.africa #### **Solution Approach** - Perform **Exploratory Data Analysis (EDA)** to understand data distribution and relationships. - Engineer meaningful features and preprocess data. - Train and optimize machine learning models. - Deploy the final model using **FastAPI** for real-time scoring. #### **How to Run the Application** 1. Install dependencies: ```bash pip install -r requirements.txt ``` 2. Start the FastAPI application: ```bash uvicorn main:app --reload ``` 3. Access the API at `127.0.0.1` for predictions. #### **Contact** Feel free to reach out for contributions or queries!