The breadth and scale of multimedia archives provides a tremendous potential for historical research that hasn't been fully tapped up to know. In this paper we want to discuss the approach taken by the History of Europe application, a demonstrator for the integration of human and machine computation that combines the power of face recognition technology with two distinctively different crowd-sourcing approaches to compute co-occurrences of persons in historical image sets. These co-occurrences are turned into a social graph that connects persons with each other and positions them, through information about the date and location of recording, in time and space. The resulting visualization of the graph as well as analytical tools can help historians to find new impulses for research and to un-earth previously unknown relationships. As such the integration of human expertise and machine computation enables a new class of applications for the exploration of multimedia archives with significant potential for the digital humanities. © 2014 Springer-Verlag.