Using satellite data in financial inclusion
This guide (PDF) by CGAP explains foundational concepts of machine learning and how financial services providers can apply those methods to leverage information contained in satellite images for the purpose of credit scoring. It focuses on smallholder farmer finance, but providers may find it useful for other applications as well. Financial services providers that see an opportunity to reach financially excluded people in rural areas can use new technology to remotely gather and analyze data on potential customers. High-quality satellite data are becoming increasingly available. By leveraging advances in machine learning (the ability of computers to analyze data quickly and at scale), providers can gain valuable insights into customers’ economic, environmental, and demographic characteristics. This guide strives to: 1) Introduce remote sensing and its potential for financial inclusion and smallholder finance. 2) Explain in simple terms how these methods work and clarify both the abilities and limits of current techniques. 3) Present use cases where computation techniques applied to satellite imagery can help organizations better serve smallholder farmers. 4) Equip organizations interested in exploring these methods with clear and actionable roadmaps that outline data prerequisites, problem scoping guidelines, and advice on getting started with research and development efforts. The guide focuses on financial services providers (FSP) that serve smallholder famers, non-FSP organisation that have smallholders farmers as customers, and development organizations and public sector. This guide can be used as a tool to help apply technologies, processes, and data analytics and machine-learning methods to improve the delivery of financial services to low-income segments.