Rich Levine, JCGS Editor
The Journal of Computational and Graphical Statistics (JCGS) is moving beyond its teenage years, riding the ever-growing wave of computational and graphical technological advances into the “Roaring Twenties,” an age of energy and fortitude mixed with growing maturity and enlightenment. To celebrate this volume, the symbols of a 20th anniversary—platinum and china—will grace the cover. Barring any delay in the production process, each issue will feature a piece highlighting the past and future of computational and/or graphical statistics, along with original research articles about a related and hot topic of the times.
The first issue features “Why Tables Are Really Much Better than Graphs,” by Andrew Gelman. The article plays off Gelman’s popular blog in a more formal journal setting, but, as the title suggests, an equally tongue-in-cheek expository aimed to rile up the troops—eliciting elaborations by graphics researchers and the statistical community—on the power of well-designed graphics.
A lively discussion ensues, leading with a historical perspective by Howard Wainer and a “just the facts ma’am” modeling perspective by Matt Briggs. Two responses follow: “Graph People vs. Table People,” by Michael Friendly and Ernest Kwan, and “Charts v. Tables: A Rematch,” by Graham Wills.
The March issue rounds out with articles about the most prominent topic covered in JCGS during its 20-year history: Bayesian computing. These works focus on the timely topics of adaptive (Markov chain) Monte Carlo, sequential Monte Carlo, model averaging, and applications in Bayesian analyses, particularly nonparametric Bayesian methods.
The second issue will focus on graphical statistics with a spread from winners of the ASA Data Expo that was organized and arranged by Hadley Wickham. The issue will also include research articles about state-of-the art visualization tools for functional and massive data and graphical innovations for complex model settings.
The third issue will feature a Shakespearean twist on a data augmentation scheme in “To Center or Not to Center: That Is Not the Question an Ancillarity-Sufficiency Interweaving Strategy (ASIS) for Boosting MCMC Efficiency,” by Yaming Yu and Xiao-Li Meng with discussion from the research groups of Jim Hobert, Gareth Roberts, Dongchu Sun, and Ying Nian Wu.
The fourth issue of Volume 20 will feature historical pieces about pervasive topics in computational and graphical statistics.
An additional high point of the 20th anniversary celebration will be an invited paper session at JSM in Miami Beach, Florida, titled “Teaching an Old Dog New Tricks: Parallel, Adaptive, and Automated Monte Carlo Methods Appearing in JCGS,” with talks by Radu Craiu, Tim Hanson, and Chris Holmes. The former two will bring to life clever manipulations of mixture models and Pólya trees, respectively, in the growing area of adaptive MCMC. The latter will highlight a recent JCGS “submission” of bringing graphics cards (GPUs) to mainstream statistical computing, an inexpensive means of performing massively parallel computing tasks—particularly simulation—right on your desktop, no supercomputer or computing cluster required.
Leave a Reply