The ability to discover specific communities of interest and to identify key concerns/opinions of particular groups or societies are important aspects of social intelligence. A number of researchers are investigating computational models of social knowledge such as emotional/empathetic state and its embodiment, social relationships, and influences on attitudes/opinions/behavior of human and computational agents (Dautenhahn 1998, 1999). In contrast, in our research we have been developing methods to process human created artifacts (e.g. documents, newsletters) and from this discover a particular class of social relationships, namely expert communities of practice. In related work, we characterize societal interests from massive information processing. This class of social knowledge is of particular relevance to establishing human communities of interest and practice which has been applied in our own research in persistent, place-based virtual environments (cvw.mitre.org, jcs.mitre.org). This paper describes two automated tools for discovering and mapping areas of social interests, groups or societies of practice. We describe XperNet, a prototype for the detection of extant or emerging communities of practice without a prior knowledge of their existence and SIAM (Social Indicator Analysis Method) which given large amounts of documents representing a community, provides indicators of common social opinion or concern.