Understanding patterns in woody plant trait relationships and trade-offs is challenging. Here, by applying machine learning and data imputation methods to a global database of georeferenced trait measurements, the authors unravel key relationships in tree functional traits at the global scale.