Financial Health Assessment for Households in Kenya

Abstract

The analysis of fiscal position represents significant in the encouragement of economic stability and development specifically to the country like Kenya. This work employs the clustering analysis technique in order to assess the Kenyan households’ financial literacy based on a financial dataset. These include transaction details, current incomes, expenditures, savings and investment practices among the clients. This paper makes use of K-means clustering technique to investigate the various clusters of households that possibly exist based on their transactional data and assess the financial characteristics and health profiles of the respective clusters. The data used was gathered from 298 distinct households across 5 counties in Kenya. Using correlation analysis and variance threshold, a feature selection method, which aimed at identifying features depicting the greatest extent of clustering tendency was also created. The clustering resulted into eight groups mainly identifiable by transaction direction, transaction value,transaction family, transaction purpose, and mode of transaction. The findings are useful in forecasting and mitigating potential risks within financial planning and management, and may improve financial wellbeing and sustainability at the household level. The project demonstrates how it is possible to implement machine learning methodologies that provide more valuable information at more efficient assessments of the financial well-being of households in Kenya.