Visualizing MDS coordinates of multiple information in color for attribute-based two-dimensional space

Abstract

Multidimensional scaling (MDS) is often used by researchers to provide a visual representation of the pattern of proximities (i.e., similarities or distances) among a set of objects normally via 2-D space map. Such representation often relates to survey questions involving subjects' multiple responses toward certain attributes in a study. However, too many subjects and attributes will produce massive output points (coordinates) which could dampen the visualization of coordinates in 2-D space map. Therefore, we propose a new way to visualize the MDS output presentation in 2-D space map by reclassifying the results according to attributes with different shapes and colours, and recalculate all the cases to view the similarity scaling according to height and distance ratio between all the output coordinates using Java programming. In this study, responses regarding preferences and reasons for choosing favorite colors were compared between subjects of different sample sizes. The purpose is to see the changes in the responses based on the new visualization of coordinates. The results had shown a marked improvement in the visual representation of the results based on different heights, shapes and colors. © 2011 Springer-Verlag.