On December 10, 2021, Leland Wilkinson, a renowned statistician and software developer, died of a massive stroke he’d suffered two days earlier. He was a youthful 77.
His daughters, Amie and Caroline, lost their doting father; his wife, Marilyn (Marny), lost her loving husband; and we lost a dear friend of 50 years.
Leland (Lee, to all who knew him) was a man of many parts. The early years he spent in a seminary gave him a particular perspective and a generosity of spirit he never lost. He was a fine pianist, which led to his affection for very grand pianos. He was also an expert programmer, which he enjoyed immensely, and a car guy. His 911 Porsche was always polished and pristine, as was Marny’s Tesla, whose remarkable torque he delighted in demonstrating to all who rode with him.
Leland Wilkinson and The Grammar of Graphics
Lee authored a beautiful book with broad impact on statistics and data science, although it has been read by only a fraction of those it has affected. The work is The Grammar of Graphics (Springer), which codifies a consistent way to represent—and, more important, to think about—statistical graphics. This is not a book to curl up with in front of a fire on a cold winter’s night; it is a work to read carefully and study, or perhaps to sample as a reference once you have digested the underlying principles. As literature, the plot is weak, but as science, the plots are better described than you’ll find anywhere else.
How can such an extensive, dense, and mathematically and computationally sophisticated tome read by relatively few have had such a wide-ranging impact? The answer is the grammar Lee laid out is the foundation of ggplot2, a graphics package created by Hadley Wickham in the R statistical computing language. If you use R, chances are you’ve used ggplot2 and thereby been touched by Lee’s work.
The Grammar of Graphics exhibits several of the many facets of Lee’s genius. Lee drew from his studies of religion, classics, psychology, music, computing, and statistics to casually drop classical references, religious quotations, and psychological insights throughout the book. They season the statistics and computer science, making reading more palatable. The breadth of his knowledge inspires awe. What other statistical computing text intelligently discusses microtonic music, quotes both the Gospel of Thomas and George Shultz, and refers important terms back to their Greek origins—written in Greek?
Lee understood both how humans perceive and how computers display, and the combination and interplay is what gives this work its power.
When I (PV) wrote as a coauthor with Lee (e.g., “Nominal, Ordinal, Interval, and Ratio Typologies are Misleading,” The American Statistician, 47(1): 65–72), his eclectic brilliance made projects more incisive and his good nature made the work more fun. As a fellow developer of statistics software, I often consulted Lee and valued his insights about the direction statistics software was headed. Technically, we were competitors in the early days of statistics software (although Lee’s Systat substantially outsold my DataDesk), but Lee was consistently generous with advice and insights. His take was clear-eyed and often quite different from what others might have thought at the time.
We regularly made it a point to lunch or dine together at least once at professional meetings so we could catch up on each other’s life, work, and views and debate the best approach to thorny problems of the day in statistical computing. It was a high point of the meeting for me. I will miss him.
Leland Wilkinson as a Modest Man and Proud Father
I (HW) am especially fond of a discussion I once had with Lee about his daughter, Amie—a world-class mathematician who has received recognition from all quarters. His pride in her accomplishments was beyond words. Lee, of course, knew a fair amount of mathematics, himself, but he explained to me it was impossible to compare the two of them—their knowledge and abilities in mathematics were simply not on the same scale. I understood him completely and asked if he considered himself like Samuel Johnson’s horse. He was familiar with Johnson’s dog (the one that walked badly on its hind legs) but didn’t know about his horse. I quoted Johnson that “a horse that can count to 10 is a remarkable horse, but it is not a remarkable mathematician.” He loved the story and reminded me of its apt characterization repeatedly thereafter.
Leland Wilkinson—Some Other Parts
Lee was also a mensch—a Yiddish word literally meaning a man but connoting a person of integrity and honor—a distinguished depiction he would sometimes bestow on others. A fine description of Lee, too, but it does not go far enough. An even more apt description comes from the Hebrew 
(transliterated as Tzadik). A Tzadik is a righteous man. Our world has far too few righteous people, and so the loss of even one is especially grievous. We grieve the loss of our friend, a true Tzadik.
Additional information
Notes on contributors
Paul Velleman
Paul Velleman is emeritus professor of statistics and statistical sciences at Cornell University. He is coauthor of six statistics textbooks (Intro Stats, Stats, Data and Models, and others, published by Pearson). He developed the DataDesk statistics software (www.datadesk.com) and DASL data and story library of teaching data sets (www.dasl.datadescription.com), both currently maintained and supported by Data description, Inc., where he serves as president and senior scientist. He is a fellow of the American Statistical Association and American Association for the Advancement of Science. He is also founder and president of Affordable Advanced Analytics (www.a3giving.org), a nonprofit company helping other small nonprofits with predictive modeling of their data. Velleman lives in Camden, Maine, with his wife and two dogs, practices Tai Chi, and sings with the a cappella group VoXX.
Howard Wainer
Howard Wainer is a statistician and author who has written Visual Revelations continuously since 1990. He has won numerous awards and is a fellow of the American Statistical Association and American Educational Research Association. His most recent book, co-authored with Michael Friendly, is A History of Data Visualization and Graphic Communication, which was published by Harvard University Press in 2021. Before this was Truth or Truthiness: Distinguishing Fact from Fiction by Learning to Think like a Data Scientist (Cambridge), which was named by the Financial Times to its “Top Six Books of 2016.”