Timeline visualization of keywords

Abstract

Visualizations of communications between actors are typically presented as actor interactions or as plots of the dates and times when the communications occurred. These visualizations are valuable to forensic analysts; however, they do not provide an understanding of the general flow of the discussed topics, which are identified by keywords or keyphrases. The ability to view the content of a corpus as a timeline of discussion topics can provide clues to when certain topics became more prevalent in the discussion, when topics disappeared from the discussion and which topics are outliers in the corpus. This, in turn, may help discover related topics and times that can be used as clues in further analyses. The goal is to provide a forensic analyst with assistance in systematically reviewing data, eliminating the need to manually examine large amounts of communications. This chapter focuses on the timeline-based visualization of keywords in a text corpus. The proposed technique employs automated keyword extraction and clustering to produce a visual summary of topics recorded from the content of an email corpus. Topics are regarded as keywords and are placed on a timeline for visual inspection. Links are placed between topics as the timeline progresses. Placing topics on a timeline makes it easier to discover patterns of communication about specific topics instead of merely focusing on general discussion patterns. The technique complements existing visualization techniques by enabling a forensic analyst to concentrate on the most interesting portions of a corpus.