Thomas Davenport, author of the best-selling Competing on Analytics, has penned a follow-up book titled Keeping Up with the Quants. Written with co-author Jinho Kim, this new title focuses on the process of analytics, and on the realities of communicating with analysts who may have more quantitative skills than yourself.
The opening chapter provides a clear-cut classification of analytic opportunities as being descriptive, predictive or prescriptive. Descriptive analytics is what was once more mundanely called reporting. Predictive analytics uses the power of statistics to predict outcomes, and prescriptive analytics strives to identify an optimal solution. Distinctions are also drawn between quantitative and qualitative methods, and some basic analytic categories such as statistics, forecasting, data mining and text mining. There is also an illuminating discussion on the nature of big data and the use of unstructured data.
Subsequent chapters deal with each of the three primary stages of the process that have been identified by the authors: framing the problem, solving the problem and presenting results. The chapter on framing the problem breaks down this task into two fairly obvious substeps: problem recognition and reviewing previous findings.
The next chapter on solving the problem is more interesting and informative. A distinction is drawn between analyses that are hypothesis-based versus those that are drive by machine-learning, where software is used to fit models to data in an automated fashion. This second approach is fundamental to currently popular data science methods. Focusing on the first approach of hypotheses-based models, the authors advocate these substeps: modeling and variable selection, data collection, and data analysis. For the data analysis stage, there are succinct descriptions of the major statistical techniques, such as causality, clustering, correlation, factor analysis, chi square, hypothesis testing and regression.
The next chapter on presenting results discusses visual analytics. The authors nicely summary some of the major tools, like scatterplots, bar charts, histograms, line graphs and pie charts, discussing the strengths of each in conveying the relationships in different types of data. As is often the case in the book, the authors combine their findings with an interesting case study; in this case, a fascinating study of divorce statistics.
Subsequent chapters turn the reader’s attention to the subject of how to manage and effectively engage quantitative analysts in the organization. The authors generally advocate the idea of not being intimidated by quantitative concepts. While this advice may seem obvious, a good number of interesting anecdotes are related to emphasize the point. The books wraps up with discussion on how business decision makers can effectively interact with analytic professionals. As expected, it is suggested that business professionals not be afraid to question quantitative results, and likewise that analytic professionals be required to understand business dynamics and implications of their work.
Overall, this is a breezy and informative presentation of the subject, filled with compelling anecdotes. Basic concepts are presented at a high enough level to prevent the reader from being immersed in unnecessary technical detail.
Keeping Up with the Quants: Your Guide to Understanding and Using Analytics
by Thomas H. Davenport and Jinho Kim
Harvard Business Review Press, June 2013
228 pages, $30.00