Machine Learning for Credit Scoring: Banking the Unbankable in Nigeria

Challenge Background

A report by Nigeria Inter-Bank Settlement System (NIBSS) shows that as of 2020, 56 million out 99.6 million adults in Nigeria are either unbanked, underbanked or financially excluded. This shows that many working adults will not have access to loans, grants and government subsidies. This could be one of the reasons for the high level of poverty in the country.

The Problem

We will build a model that houses the details of people that are unbanked in Nigeria.

The aim is to build tools that are accessible to every adult. They can subsequently register their details. Their details will be collected and stored in the system. Any financial institution, government, or NGO can log onto the system to access the creditworthiness of an individual. This will help them in decision-making about whether to avail the loan to the user or not. 

The project results will be made open source. The solution will help the country alleviate poverty by making credit facilities available to everyone, even the unbanked. It will also help the government in development planning and allocation of subsidies to the people. The project can be further developed to be integrated into the systems of financial institutions and governments.

Goal of the Project

The achievable goal of the project is:

• Determine the credit worthiness of an unbanked customer with alternate and traditional credit scoring data and methods.
• Build the model in such a way that it is accessible to anyone from an app or a website.
• Serve as a database for the unbanked in Nigeria.

Project Timeline

What you'll learn

1. Web Scraping

2. Machine Learning

3. Data Analysis

4. Data Preparation