In fields ranging from security to systems biology to business, the ability to collect data far outstrips the ability to analyze the data that have been collected. The science of visual analytics is meant to enable analysts of all types to overcome information overload so that they can detect the expected and discover the unexpected from massive, dynamic, conflicting, and incomplete information while rigorously adhering to privacy and security laws and policies. During the past year, a panel of research leaders from academia, industry government, the national laboratory system developed a five year research and development agenda for the field of visual analytics. This article presents a concise summary of that agenda. © 2006 IEEE.