June 15. 2007 by Del.

Steve Gibson of grc.com and Security Now guided me to SpaceMonger - a disk / file visualization tool. Think in terms of size, age, and other file attributes - then map multiple variables of a hierarchy in a map with size and color to differentiate. The results are stunning visuals that let you immediately hone in on relevant criteria. Over the last few years similar maps of the stock market with industry and company information have become popular.

But now the best part! All of this technology is based on research conducted at the University of Maryland. Documentation and flash video tutorials are online and available!
Treemap is a space-constrained visualization of hierarchical structures. It is very effective in showing attributes of leaf nodes using size and color coding. Treemap enables users to compare nodes and sub-trees even at varying depth in the tree, and help them spot patterns and exceptions.

The research team has also developed tools (free for academic use) to create treemap. The current version is 4.1.1 and is available at http://www.cs.umd.edu/hcil/treemap/ .

Posted in Visualization, DataAnalysis, statistics | Print | No Comments »

May 26. 2004 by Del.

Excellent summary and bibliography of sample size rules of thumb for statistical analysis. ddv
Statistical Rules of Thumb: What We Don’t Want to Forget About Sample Sizes - Psi Chi: “Statistical Rules of Thumb: What We Don’t Want to Forget About Sample Sizes

by Carmen Wilson VanVoorhis and Betsy Levonian Morgan - University of Wisconsin-La Crosse

In this article we highlight the statistical rules of thumb guiding the selection of sample sizes for detecting differences, associations, chi-square, and factor analyses.

Cohen (1990), as an eminent psychometrician, had the experience and depth to write an article entitled “Things I Have Learned (So Far).” However, if you are a student grappling with learning the complexities of research, you may want an article entitled “What We Need to Know.” If you are faculty aiding undergraduate projects, you may wish to join us in focusing on the utilitarian principles of research design that we “don’t want to forget.” Concerned by the ease of computer programs that tend to lure a person into hedging on the tried-and-true of statistical theory, we were motivated to reestablish some of the solid heuristics on sample size that guide our research. Indeed, issues regarding power and effect size and their implications for sample size selection are pressing themes in research design and statistical implications (Kraemer & Thiemann; 1987; Wilkinson, 1999; Wolins, 1982). Consequently, we offer this concise review and summary of the “rules of thumb” regarding sample size in the spirit of improving research design and aiding our students in their research endeavors.”

Posted in statistics | Print | No Comments »