Secondary Logo

Analysis of Large and Complex Data

Wanderer, Jonathan P., MD, MPhil

doi: 10.1213/ANE.0000000000002127
Books, Multimedia, and Meeting Reviews: Book Review
Free

Department of Anesthesiology, Vanderbilt University Medical Center, Nashville, Tennessee, jonathan.p.wanderer@vanderbilt.edu

When Analysis of Large and Complex Data hits your desk, it will be with a thud. A satisfying thud, yet a thud nonetheless. While the existence of big data in the field of anesthesiology has been previously debated1—and on all accounts, the presence of billions of rows and many columns does not quite translate to the functional specification of velocity, volume, and variety—there is no debate that this rich collection of papers from the Second European Conference on Data Analysis represents a sizable contribution to a spectrum of scientific endeavors. This collection of papers is broken into a series of themes, starting with big data, clustering, and regression, and continuing with focused chapters in Data Analysis that range from Data Analysis in Marketing through Data Analysis in Library Science. While the initial few chapters are of general utility, the focus that subspecialization provides quickly becomes apparent.

In the several hundred pages that intervene between Marketing and Library Science lies Data Analysis for Medicine and Life Sciences. While the chapter title is inviting, on both a relative and absolute basis, of the papers on offer, the only one that might appeal to the clinician is “Estimating Age- and Height-Specific Percentile Curves for Children Using GAMLSS in the IDEFICS Study.” This paper describes work related to generating growth curves and describes the merits and technical details required to turn tens of thousands of individual growth records into modeled population-level data. Turning the page beyond this paper yields topics of a more esoteric nature that might appeal to the computer science-trained clinician, with “An Ensemble of Optimal Trees for Class Membership Probability Estimation” and the related “Ensemble of Subset of k-Nearest Neighbors Models for Class Membership Probability Estimation.” Any readers adventuresome enough to flip beyond these 3 papers in the medicine section will run headlong into Data Analysis in Musicology as a result of the vagaries of alphabetical organization, where they will encounter engaging entrées on the character of music. A description of these musicology papers is out of the scope of this book review.

While there are a number of accessible texts on statistical analysis that are to be recommended for the anesthesiologist-turned-data-scientist, this is unfortunately not one of them. There is no intentional effort made to connect each paper, or chapter, together, as this tome represents a collection of papers presented at a data science meeting rather than an attempt to knit together a set of diverse topics to appeal to a broad group of readers. Several of the papers within the clustering, classification, and regression chapters are worth reading in their entirety. In particular, “Reviewing Graphical Modelling of Multivariate Temporal Processes” and “The Weight of Penalty Optimization for Ridge Regression” are compelling reads. However, on the balance, the content contained within these specific papers would more likely be better accessed through the traditional peer-reviewed publishing process. The context that this weighty book provides is not worth the additional overhead. The reader would be better off preserving both shelf space and dollars by pursuing papers on these same topics on their own individual merit.

Analysis of Large and Complex Data is a useful book for those who are very well read in sophisticated data analysis across variety of disciplines and are looking for a book that captures state-of-the-art scientific papers across a wide cross-section. For the rest of us who might be looking for a focused view on analysis of complex health care datasets or a thorough and cohesive review of complex data analysis in general, however, our efforts would be better rewarded by spending time with a different resource.

Jonathan P. Wanderer, MD, MPhil

Department of Anesthesiology

Vanderbilt University Medical Center

Nashville, Tennessee

jonathan.p.wanderer@vanderbilt.edu

Back to Top | Article Outline

REFERENCE

1. Levin MA, Wanderer JP, Ehrenfeld JM. Data, big data, and metadata in anesthesiology. Anesth Analg. 2015;121:1661–1667.
© 2017 International Anesthesia Research Society