HexBoard: Conveying Pairwise Similarity in an Incremental Visualization Space

We introduce the HexBoard space and visualization tool, that provides an incremental visualization space that conveys pairwise similarity among all neighboring elements from a dynamic data set. It is a significant evolution on our previous work, incBoard, a chess board analogy for displaying (projecting) objects from a dynamic set on a 2D space, considering their similarity in a higher dimensional space. HexBoard relays on hexagons to represent data items: the only regular polygons that provide a regular tessellation of the Euclidean plane and where neighboring elements always share an edge. These edges are then easily manipulated to convey pairwise similarity information, thus overcoming a serious limitation of the previous solution, while preserving all advantages of incBoard (no occlusion, coherent disposition of elements, inherently incremental solution and low computational cost). This paper introduces HexBoard, discusses its potential and illustrates its application with examples.