As industries increasingly rely on innovation and knowledge workers, interaction and social networks have once again emerged as a primary purpose for working together in the place called office. The purposes of this thesis are (1) to investigate the roles of physical setting, especially on building and site scale, and its influence to foster communication and interaction in large organizations, and (2) to explore analytical tools and techniques to verify the performance of workplace spatial attributes and interaction patterns. Employing the archival data of interaction patterns and physical data acquired from satellite images and GIS, This thesis investigates the relationship between spatial attributes and interaction emerged in four corporate campuses, including Goldman Sachs, Sprint, Sun Microsystems, and Toyota Motor Sale. The analysis reveals that there are very low correlations for both relationships between floor plate area and interaction, and relationships between floor plate ratio and interaction. In addition, there is no predictable trend between the average travel time between building and the average interaction frequency. There is a moderate correlation between actual travel time and self-reported or perceived travel time. The thesis also suggests that the self-reported travel time tends to be under-estimated by the employees. The actual travel time, however, has more impact on interaction in compact campuses than in dispersed campuses.