Praise for the First Edition
"The main strength of this book is that it provides a unified framework of graphical tools for data analysis, especially for univariate and low-dimensional multivariate data. In addition, it is clearly written in plain language and the inclusion of R code is particularly useful to assist readers’ understanding of the graphical techniques discussed in the book. … It not only summarises graphical techniques, but it also serves as a practical reference for researchers and graduate students with an interest in data display." -Han Lin Shang, Journal of Applied Statistics
Graphics for Statistics and Data Analysis with R, Second Edition, presents the basic principles of graphical design and applies these principles to engaging examples using the graphics and lattice packages in R. It offers a wide array of modern graphical displays for data visualization and representation. Added in the second edition are coverage of the ggplot2 graphics package, material on human visualization and color rendering in R, on screen, and in print.
Features
- Emphasizes the fundamentals of statistical graphics and best practice guidelines for producing and choosing among graphical displays in R
- Presents technical details on topics such as: the estimation of quantiles, nonparametric and parametric density estimation; diagnostic plots for the simple linear regression model; polynomial regression, splines, and locally weighted polynomial regression for producing a smooth curve; Trellis graphics for multivariate data
- Provides downloadable R code and data for figures at www.graphicsforstatistics.com
Kevin J. Keen is a Professor of Mathematics and Statistics at the University of Northern British Columbia (Prince George, Canada) and an Accredited Professional StatisticianTM by the Statistical Society of Canada and the American Statistical Association.
List of Figures
List of Tables
Preface to the First Edition
Preface to the Second Edition
Acknowledgments
I Introduction
The Graphical Display of Information
Introduction
Learning Outcomes
Know the Intended Audience
Principles of Effective Statistical Graphs
The Layout of a Graphical Display
The Design of Graphical Displays
Graphicacy
The Grammar of Graphics
Graphical Statistics
Conclusion
Exercises
II A Single Discrete Variable
Basic Charts for the Distribution of a Single Discrete Variable
Introduction
Learning Outcomes
An Example from the United Nations
The Dot Chart
The Bar Chart
Definition
Pseudo Three-Dimensional Bar Chart
The Pie Chart
Definition
Pseudo Three-Dimensional Pie Chart
Recommendations Concerning the Pie Chart
Conclusion
Exercises
Advanced Charts for the Distribution of a Single Discrete Variable
Introduction
Learning Outcomes
The Stacked Bar Chart
Definition
The Stacked Bar Plot Versus the Bar Chart and the Pie Chart
The Pictograph
Definition
The Pictograph Versus the Dot Chart and the Bar Chart
Variations on the Dot and Bar Charts
The Bar-Whisker Chart
Dot-Whisker Chart
Frames, Grid Lines, and Order
Frame
Grid Lines
Order
Conclusion
Exercises
III A Single Continuous Variable
Exploratory Plots for the Distribution of a Single Continuous Variable
Introduction
Learning Outcomes
The Dotplot
Definition
Variations on the Dotplot
The Stemplot
Definition
The Boxplot
Definition
Variations on the Boxplot
The EDF Plot
Definition
The EDF Plot as a Diagnostic Tool
Conclusion
Exercises
Diagnostic Plots for the Distribution of a Continuous Variable
Introduction
Learning Outcomes
The Quantile-Quantile Plot
The Probability Plot
Estimation of Quartiles and Percentiles∗
Estimation of Quartiles
Estimation of Percentiles
Conclusion
Exercises
Nonparametric Density Estimation for a Single Continuous Variable
Introduction
Learning Outcomes
The Histogram
Definition
A Circular Variation on the Histogram: The Rose Diagram
Kernel Density Estimation∗
Spline Density Estimation∗
Choosing a Plot for a Continuous Variable∗
Conclusion
Exercises
Parametric Density Estimation for a Single Continuous Variable
Introduction
Learning Outcomes
Normal Density Estimation
Transformations to Normality
Pearson’s Curves∗
Gram-Charlier Series Expansion∗
Conclusion
Exercises
IV Two Variables
Depicting the Distribution of Two Discrete Variables
Introduction
Learning Outcomes
The Grouped Dot Chart
The Grouped Dot-Whisker Chart
The Two-Way Dot Chart
The Multi-Valued Dot Chart
The Side-by-Side Bar Chart
The Side-by-Side Bar-Whisker Chart
The Side-by-Side Stacked Bar Chart
The Side-by-Side Pie Chart
The Mosaic Chart
Conclusion
Exercises
Depicting the Distribution of One Continuous Variable and One Discrete Variable
Introduction
Learning Outcomes
The Side-by-Side Dotplot
The Side-by-Side Boxplot
The Notched Boxplot
The Variable-Width Boxplot
The Back-to-Back Stemplot
The Side-by-Side Stemplot
The Side-by-Side Dot-Whisker Plot
The Trellis Kernel Density Estimate∗
Conclusion
Exercises
Depicting the Distribution of Two Continuous Variables
Introduction
Learning Outcomes
The Scatterplot
The Sunflower Plot
The Bagplot
The Two-Dimensional Histogram
Definition
The Levelplot
The Cloud Plot
Two-Dimensional Kernel Density Estimation∗
Definition
The Contour Plot
The Wireframe plot
Conclusion
Exercises
V Statistical Models for Two or More Variables
Simple Linear Regression: Graphical Displays
Introduction
Learning Outcomes
The Simple Linear Regression Model
Definition
The Scatterplot
The Sunflower Plot
Residual Analysis
Definition
Residual Scatterplots
Depicting the Distribution of the Residuals
Depicting the Distribution of the Semistandardized Residuals
Influence Analysis
Definition
Matrix Notation for the Simple Linear Regression Model
Depicting Standardized Residuals
Depicting the Distribution of Studentized Residuals
Depicting Leverage
Depicting DFFITS
Depicting DFBETAS
Depicting Cook’s Distance
Influence Plots
Conclusion
Exercises
Polynomial Regression and Data Smoothing: Graphical Displays
Introduction
Learning Outcomes
The Polynomial Regression Model
Splines
Locally Weighted Polynomial Regression
Conclusion
Exercises
Visualizing Multivariate Data
Introduction
Learning Outcomes
Depicting Distributions of Three or More Discrete Variables
The Sinking of the Titanic
Thermometer Chart
Three-Dimensional Bar Chart
Trellis Three-Dimensional Bar Chart
Depicting Distributions of One Discrete Variable and Two or More Continuous Variables
Anderson’s Iris Data
The Superposed Scatterplot
The Superposed Three-Dimensional Scatterplot
The Scatterplot Matrix
The Parallel Coordinates Plot
The Trellis Plot
Observations of Multiple Variables
OECD Healthcare Service Data
Chernoff’s Faces
The Star Plot
The Rose Plot
The Multiple Linear Regression Model
Definition
Modeling Perch Mass
Residual Scatterplot Matrix
Leverage Scatterplot Matrix
Influence Plot
Partial-Regression Scatterplot Matrix
Partial-Residual Scatterplot Matrix
Summary of the Model for Perch Mass
Conclusion
Exercises
VI Appendices
Human Visualization
Introduction
Learning Outcomes
Optics
Introduction
Geometrical Optics
The Light Spectrum
Anatomy of the Human Eye
The Perception of Colour
Graphical Perception
Weber’s Law
Stevens’s Law
The Gestalt Laws of Organization
Kosslyn’s Image Processing Model
Conclusion
Exercises
Color Rendering
Introduction
Learning Outcomes
RGB and XYZ Color Spaces
HSL and HSV Color Spaces
CIELAB and CIELUV Color Spaces
HCL Color Space
CMYK Color Space
Displaying Color in R
Saving Color Documents from R
Conclusion
Exercises
Bibliography
Index
Biography
Kevin J. Keen is a Professor of Mathematics and Statistics at the University of Northern British Columbia (Prince George, Canada) and an Accredited Professional StatisticianTM by the Statistical Society of Canada and the American Statistical Association.
"A leading expert wrote the book. The book is an exposition of statistical methodology that focuses on ideas and concepts and makes extensive use of graphical presentation, but readers should have some prior experience of statistical methodology. The chapters also contain many exercises with solutions and hints presented in the Appendix. The R codes are available for download on the website. The book presents data and Programmes to replicate the models developed, offers new methods that are ready to use, and explores graphical statistics in its entirety from the fundamentals of modern methods. The book is also a complete reference manual and should be considered a must-have companion for the interested advanced audience."
~International Society for Clinical Biostatistics
". . . this is a book I can recommend for consideration in a course or as a course supplement. It is generally clear and well-written, and the statistical aspects of some of these methods are explained in sufficient detail to put these in context."
~Michael Friendly, Journal of Agricultural, Biological, and Environmental Statistics
