Drawing Social Networks Using Area-Labeling Rectangular Cartograms

Abstract

People are linked together by social networks, whose complexity depends upon a lot of factors. To our understanding, most visualization interfaces for social networks have not been designed for reflecting their factors so far. As a result, this paper tries to solve such a problem by rectangular cartogram, which is a kind of geographical visualization interface using rectangles to represent regions in a map. Besides the relative position of each rectangle can reflect actual geographical related positions, one of the main features of rectangular cartograms is to use the area size or the shape of each rectangle to reflect the information of its corresponding region, e.g., the population in that region. This paper proposes a layout approach for rectangular cartograms with area labeling for social networks, in which each region has a minimum-width constraint for accommodating a text label. To satisfy the practical use for social networks, we apply a genetic algorithm to finding the minimum-width area-labeling rectangular cartogram under some constraints. By doing so, we can visualize the labeling text on each rectangle and observe the information represented by its area size or shape at the same time. Furthermore, the proposed approach is applied to visualizing the distribution of the Facebook popularity of an enterprise in Taiwan. From the cartogram, the text label on each region can be read directly and the relation among regions as well as their popularity can be visualized at the same time, so that the enterprise can improve the regions with poor popularity by the help from the regions with high popularity.