StoryGem: Voronoi treemap Approach for Semantics-Preserving Text Visualization

Word cloud use is a popular text visualization technique that scales font sizes based on word frequencies within a defined spatial layout. However, traditional word clouds disregard semantic relationships between words, arranging them without considering their meanings. Semantic word clouds improved on this by positioning related words in proximity; however, still struggled with efficient space use and representing frequencies through font size variations, which can be misleading because of word length differences. This paper proposes StoryGem, a novel text visualization approach that addresses these limitations. StoryGem constructs a semantic word network from input text data, performs hierarchical clustering, and displays the results in a Voronoi treemap. Furthermore, this paper proposes an optimization problem to maximize the font size within the regions of a Voronoi treemap. In StoryGem, word frequencies map to area sizes rather than font sizes, allowing flexible text sizing that maximizes use of each region's space. This mitigates bias from varying word lengths affecting font size perception. StoryGem strikes a balance between a semantic organization and spatial efficiency, combining the strengths of word clouds and treemaps. Through hierarchical clustering of semantic word networks, it captures word semantics and relationships. The Voronoi treemap layout facilitates gapless visualization, with area sizes corresponding to frequencies for clearer representation. User study across diverse text datasets demonstrate StoryGem's potential as an effective technique for quickly grasping textual content and semantic structures.