Statistically sound distribution plots in Excel

Abstract

Excel is the most widespread and the most powerful general-purpose spreadsheet software, but it is not popular with statisticians. Nevertheless, as a natural means for organising, displaying and analysing large amounts of data, spreadsheets keep gaining importance in statistical education and practice. Aiming at improving such practice rather than fruitlessly and indiscriminately condemning it, the paper provides general considerations on the topic, pointers to the huge body of relevant literature and software, and several concrete examples of data visualisation in Excel in the sense of univariate, bivariate and multivariate distribution plotting. Original and improved Excel solutions for producing dot-density plots, dot plots, stem-and-leaf plots, windowgrams, coplots and parallel coordinates plots are presented, as well as for performing the Box-Cox transformation. Additionally, further possibilities opening with the forthcoming Excel 2007 version, use of various commercial and freeware add-ins, and integration of Excel with statistical software are discussed.