NOVEL METHODS IN THE VISUALIZATION OF TRANSITIONAL PHENOMENA
Download the Paper Forward to a colleague
Abstract:
History matrix visualization is presented as an exploratory tool to enable the exploration of survival data for groups of several individuals longitudinally followed through multiple states. History matrix visualizations have a tremendous data density and depict the data in its entirety as well as allowing for individual and group level visual and numerical summarizations. The method allows for an interactive graphing experience and produces graphs that are inherently linked. Herein an implementation of the history matrix visualization method is demonstrated using the sleep data of 59 matched pairs of subjects with and without sleep apnea participating in the Sleep Heart Health Study. Results showed that those with sleep apnea experience more periods of short-term and long-term wakefulness during the night and more transitions between sleep stages than subjects without sleep apnea. Marked differences were also found in the temporal evolution of rapid eye movement (REM) sleep between the two groups.
Subject Area:
Longitudinal Data Analysis and Time Series
Suggested Citation:
Bruce Swihart, Brian Caffo, Matthew Strand, and Naresh M. Punjabi, "NOVEL METHODS IN THE VISUALIZATION OF TRANSITIONAL PHENOMENA" (April 2007). Johns Hopkins University, Dept. of Biostatistics Working Papers. Working Paper 141.
http://www.bepress.com/jhubiostat/paper141