Dynamic and Interactive Reproducible Statistical Documents

• Activity Number: 190
• Type: Invited
• Date/Time: Tuesday, August 13, 2002 : 10:30 AM to 12:20 PM
• Sponsor: Amstat Online
• Author(s): Duncan Lang*+
• Affiliation(s): Bell Labs, Lucent Technologies

Abstract

Traditionally, the process of analyzing data, performing simulations and general statistical computations has been done separately from the preparation of documents and reports in which they are presented. The output from statistical systems is frequently inserted into Word and LaTeX documents manually and in ad hoc ways. The documents are not reproducible or verifiable. Additionally, they are static and ``closed'' in that the reader cannot update the contents of the document directly or by introducing new or additional data. We have developed two tools that integrate statistical systems, both R and Omegahat, into both the document preparation and display process. The first extends an XSL (the eXtensible Stylesheet Language) translator--Xalan--to allow functions in these systems to be called when dynamically generating content that is described using XML (the eXtensible Markup Language), and transformed into PDF, Postscript, HTML, etc. This provides a rich form of reproducibility and a formal way for integrating output from statistical systems directly into documents. It exploits the general popularity of XML and its cousin, XSL, for marking up and transforming documents.