Title: 100+ years of graphs of the Titanic data
Authors: Juergen Symanzik - Utah State University (United States) [presenting]
Michael Friendly - York University - Psychology Department (Canada)
Ortac Onder - York University - Schulich School of Business (Canada)
Abstract: Many readers are likely familiar with the stories of the tragic fate of passengers and crew of the RMS Titanic upon her fatal collision with an iceberg and her sinking in the early hours of April 15, 1912 on her maiden voyage to New York City. Little known is the fact that the first graphical summary of the initial survivor data appeared in The Sphere, a British newspaper, on May 4, 1912. The public inquiries that followed produced detailed data sets that have been widely used to illustrate graphical and statistical methods for quite some time. Numerous follow-up studies have used a wide variety of graphical representations related to the Titanic disaster, published in statistics, information visualization, and social sciences venues. It seemed timely to survey the variety of graphical methods used for these data sets over the last century. Graph types used to portray the Titanic data include: bar charts, mosaic plots, double-decker plots, parallel set plots, Venn diagrams, balloon plots, nomograms, and tree diagrams to name only a few. Three questions are addressed: (i) What types of graphs have been used for the Titanic data in the last 100+ years? (ii) Are some of these graphs unique and provide additional insights that are hard to obtain from the other ones? (iii) Do the graphs used differ by genre or scientific discipline?