Usability Views Article Details  home | timeline | recent | popular | e-reports | userati | books | about  A Rank-by-Feature Framework for Interactive Exploration of Multidimensional Data (08 Dec 2004) The rank-by-feature framework enables users to a systematic approach to understanding the dimensions and finding important features using axis-parallel 1D and 2D projections of multidimensional data sets. Users begin by selecting a ranking criterion and then can see the ranking for all 1D or 2D projections. They can select high or low ranked projections and view them rapidly, or sweep through a group of projections in an ordered manner. The score overview provides a visual summary that helps users identify extreme values of criteria such as correlation coefficients or uniformity measures. Information visualization principles and techniques such as dynamic query by item sliders, combined with traditional graphical displays such as histograms, boxplots, and scatterplots played a major role in the rank-by-feature framework. Article URL: ftp://ftp.cs.umd.edu/pub/hcil/Reports-Abstracts-Bibliography/2004-31html/2004-31.htm Read 41 more articles from University Of Maryland sorted by date, popularity, or title. Next Article: Interactive Pattern Search in Time Series     This site is a labour of love built by Chris McEvoy