Abstract
Literate Statistical Practice (LSP, Rossini, 2001) describes an approach for creating self-documenting statistical results. It applies literate programming (Knuth, 1992) and related techniques in a natural fashion to the practice of statistics. In particular, documentation, specification, and descriptions of results are written concurrently with writing and evaluation of statistical programs. We discuss how and where LSP can be integrated into practice and illustrate this with an example derived from an actual statistical consulting project. The approach is simplified through the use of a comprehensive, open source toolset incorporating Noweb, Emacs Speaks Statistics (ESS), Sweave (Ramsey, 1994; Rossini, et al, 2002; Leisch, 2002; Ihaka and Gentlemen, 1996). We conclude with an assessment of LSP for the construction of reproducible, auditable, and comprehensible statistical analyses.
Disciplines
Bioinformatics | Computational Biology | Numerical Analysis and Computation
Suggested Citation
Rossini, Anthony and Leisch, Friedrich, "Literate Statistical Practice" (March 2003). UW Biostatistics Working Paper Series. Working Paper 194.
https://biostats.bepress.com/uwbiostat/paper194
Included in
Bioinformatics Commons, Computational Biology Commons, Numerical Analysis and Computation Commons