The document proposes a graph traversal approach called Dedalo to explain patterns in Linked Data using background knowledge from the graph. Dedalo uses an A* search algorithm to iteratively build a graph by dereferencing URIs, find explanation paths between initial nodes and entities in the graph, and build explanations by evaluating paths and final nodes. It finds the highest quality explanations for patterns like countries where men are more educated than women. Future work aims to improve Dedalo by addressing biases from incomplete data and combining multiple explanations.