Skip to Main Content U.S. Department of Energy
Knowledge Discovery & Informatics


Retrospective Analysis of Communication Events

The Retrospective Analysis of Communication Events (RACE) prototype allows analysts to review communication data (e.g., e-mail traffic, instant messaging information, meeting transcripts, chat data, search engine interactions, etc.). Analysts can concentrate on a single modality or fuse various modalities to understand an underlying message or rationale. A three-screen "triage" process takes the analyst from a visualization of a large communications corpus, through a combined sequence view where traffic of interest can be reviewed in context, to a detailed view where a single communications event can be analyzed in detail. In conjunction with the Glass Box Analytical Environment (GBAE), the RACE tool also can help in the analysis of human interactions with intelligent components such as expert systems, next-generation search engines, and reasoners.

Click for a larger version.

Computer-mediated communications have fundamentally changed the way that individuals around the world communicate. Many people have daily professional interactions with individuals they have never met face to face. Increasingly, people are choosing to use new communication technologies for their interactions rather than traditional face-to-face contact. Although the prevalence of these new technologies at work and at home has enabled new means of collaboration, it has also provided employers with new security challenges.

The efficient analysis of communications data is important for a number of reasons.

  • Only by truly understanding how individuals use these new technologies will system designers be able to build tools and technologies to complement them. The RACE tool provides a test bed for research into anthroposemiotics (i.e., human communication), thus allowing researchers to uncover how we communicate with ourselves (i.e., intrapersonal communication), with others (i.e., interpersonal communication), within groups (i.e., group-dynamic communication), and across cultures (i.e., cross-cultural communication).
  • There is a need to be able to analyze communication data after serious events. The Enron scandal, the 2003 failure of the electric power grid in a large portion of the country, and the Hurricane Katrina disaster each produced large sets of communication data that needed to be analyzed to ascertain the roles and responsibilities of organizations and individuals involved. A tool such as RACE can make this daunting task much more manageable.
  • RACE can be an invaluable tool that individuals and teams can use to track their own communications. Major topics discussed, social factors such as positive and negative feelings, and group dynamics are all easily accessible through use of the RACE tool, potentially revealing new information on how we work and helping identify areas that can be improved.


  • Cowell AJ, JN Haack, and DW McColgin.  2006.  "Retrospective Analysis of Communication Events - Understanding the Dynamics of Collaborative Multi-Party Discourse."  In Proceedings of the Analyzing Conversations in Text and Speech (ACTS) Workshop at HLT-NAACL, New York City, New York, June 2006, pp. 62-70.  Association for Computational Linguistics, East Stroudsburg, PA. 
  • Cowell AJ, ML Gregory, JR Bruce, JN Haack, DV Love, SJ Rose, and AH Andrew.  2006.  "Understanding the Dynamics of Collaborative Multi-Party Discourse."  Information Visualization 5(4):250-259.  doi:10.1057/palgrave.ivs.9500139

Knowledge Discovery & Informatics at PNNL

Scientific Foundations



Andrew J. Cowell, Ph.D.