ABSTRACT
This chapter introduces all of the concepts and discusses the applicability to real world problems. It describes visualization techniques to derive data clusters in a completely interactive and visual way and outlines in detail techniques bases on scatterplots as well as parallel coordinates. The chapter shows approaches to visually steer automatic, unsupervised clustering algorithms and considers the interactive comparison of several readily derived clusterings. It suggests that visual inspection of clustering results and sensemaking. Clustering algorithms group data objects together based on some notion of distance or similarity. Many algorithms have been proposed to formalize the concept of a cluster and to automatically detect such clusters in large sets of data objects. Automated clustering algorithms are difficult to understand for nonexperts. Using interactive brushing the user can color different points to mark visually detected clusters. Several approaches have been suggested that address the problems and especially enhance the visual recognition of clusters in parallel coordinate displays.
