As many countries across meso-America come closer to elimination goals, understanding the transmission potential in areas where little to no transmission has occurred in the recent years becomes increasingly important to guide elimination and prevent local re-introduction of transmission. Honduras has seen significant recent declines in malaria; particularly Islas de la Bahia. Supporting the national program, in collaboration with CHAI, the goal of our analysis was to understand the underlying process that have determined the locations in which cases have appeared in the last three years to enable development of a risk map depicting geographical variation in underlying transmission potential. A Poisson point process model was used in conjunction with a machine learning tool (boosted regression trees) to understand the pattern of case occurrence in Islas de la Bahia. The resulting predictions were made at 100m resolution to provide fine-scale granularity in understanding the varying transmission risks across the small islands.