Livelihood mapping can play a crucial role in understanding and addressing poverty and development challenges. This study presents a novel methodology for high-resolution livelihood mapping using openly available geospatial data and a Kohonen Self-Organizing Map clustering algorithm. The study focuses on Kenya and aims to address three research objectives: (1) determine whether publicly available Earth Observation data can be effectively used to identify different livelihood zones in an asset-based approach, (2) develop a data-driven workflow for delineation of different livelihood zones within Kenya and (3) explore the value of this methodology for the future of socio-economic data collection. This study uses five different geospatial datasets to effectively create 25 clusters of livelihoods at the sub-county level in Kenya. Comparison with expert-created livelihood zones from 2011 and Demographic and Health Surveys (DHS) data further validate the approach's ability to differentiate livelihood patterns. The methodology also exhibits a significant correlation with wealth measures, providing deeper insights into poverty dynamics and offering potential for informing development strategies. Despite limitations this methodology presents a cost-effective and timely method of regularly updating livelihood mapping. By bridging the gap between high-resolution and large-scale livelihood mapping, this study contributes to the advancement of socio-economic research and poverty alleviation efforts.