I recently posted about Stacked Bar Chart Alternatives and included some rather cluttered charts with markers and lines, which I referred to as “Dot Plots”. I was called to task by none other than Naomi Robbins, author, researcher, consultant, and renowned expert in the field of data presentation graphics. Briefly, Naomi recoiled at the thought of my ugly graphics being grouped with the clean and effective Dot Plots that she has written about extensively. I requested that she follow up her comments with a blog post, and she has graciously provided the following article.
Some Comments on Dot Plots by Naomi Robbins
Jon Peltier posted a blog entitled “Stacked Bar Chart Alternatives” which generated active discussion. One of the alternative charts that Jon presented and then rejected as being too cluttered was called a “dot plot”. Since I am a major fan of dot plots, I cringed when I saw this and was further disturbed when I saw other blogs about stacked bar chart alternatives picking up on this. This post presents trellis dot plots of the data from the stacked bar chart and provides some comments and references about dot plots.
Bill Cleveland and his colleagues at Bell Labs introduced dot plots after conducting carefully controlled experiments on human perception as it relates to decoding information from graphs . Cleveland uses dot plots extensively in his two books: The Elements of Graphing Data and Visualizing Data. I also make frequent use of dot plots in Creating More Effective Graphs. Stephen Few invited me to write an article about dot plots for the February 2006 issue of the B-Eye newsletter . However, you will not see a figure that resembles the tangled mess of the example in the Peltier Tech Blog labeled dot plots in any of the references above.
These two figures show the type of graphs that appear in Cleveland and my works. I show them more to explain what we mean by dot plots and trellis dot plots than to add another alternative to the stacked bar chart. As many of the comments point out, the best chart depends on the purpose the designer has in creating the chart. I might choose a trellis chart if my point was “Here are the results of the survey” while I might prefer Jeff’s version if I represented Tableau and wanted to show its capabilities. I am a big believer in showing data more than one way since each presentation highlights a different aspect of the data.
Of course, I could have used color but I wanted to emphasize the fact that these charts do not depend on color the way that the stacked bar chart does. There still are many publishers that don’t allow color and many people reproduce charts on black and white copiers.
These figures were created using S-Plus 6.2, an old version of S-Plus. If any of you are R users, contact me at naomi at nbr-graphs dot com if you’re willing to help me figure out what’s wrong with my R code (the grid lines don’t show even though I copied and pasted code from another figure where they do show.)
Dot Plots or Bar Graphs
Both bar charts and dot plots are used when there are both categorical and quantitative variables. Which of these should we use?
Stephen Few says :
Because the endpoints of bars encode quantitative values, points (e.g., dots) at the same locations as the bars’ endpoints could replace the bars and convey the same meaning. So why use bars at all? Bars do one thing extremely well: due to their visual weight, they stand out so clearly and distinctly from one another that they do a great job of representing individual values discretely. 
I repeatedly have said that bar charts get more cluttered than dot plots. At first glance it would seem that we are in disagreement. However, if you look more closely, you will notice that many of Few’s examples have four or five data points such as the four quarters of a year or the four regions of the country while many of my examples have around 50 data points such as the 50 states. I have seen documents by Few where he uses dot plots with more data points and I will readily admit to using bar charts if I only have a handful of data to plot. Therefore, as with most issues, Steve and I agree on this one. See  for a discussion of advantages of dot plots over bar charts. Dot charts are appropriate when it is desirable not to include zero on the axis, error bars show up better on dot plots, and logarithmic scales work better with dot plots.
Dot plots and lines
You will notice that Jon used lines to connect the dots of the different vendors while my dots are not connected. In general, I feel that it is much more important to connect the dots to their labels than to other data points, although with as many series as Jon used, the lines were needed to distinguish the data series. Also, the pattern that you get by connecting the points is arbitrary; a different ordering of the labels would produce a different pattern. Obviously, I do not like connecting the points. However, I am well aware that many graph experts I respect disagree with me on this issue. One example is Dan Carr who connects the points on his linked micromaps (see page 136 of ).
Dot plots and Excel
I am unaware of utilities to draw dot plots in Excel that were available before Creating More Effective Graphs was published. I asked Ken Klein, an Excel user, to write a macro for me to provide to readers of the book. Kelly O’Day wrote one of the first Amazon reviews of the book and posted instructions on his Web site for creating graph forms that I recommend in Excel. Then Charlie Kyd and Jon provided instructions and/or utilities to produce these plots. Links to their sites are available at http://www.nbr-graphs.com/trainframe.html. Thank you, Jon, for providing this much needed service to Excel users.
1. Cleveland, William S. and Robert Mc Gill. 1984. “Graphical Perception: Theory, Experimentation, and Application to the Development of Graphical Methods.” Journal of the American Statistical Association 79:531-554.
2. Cleveland, William S. 1994. The Elements of Graphing Data. Revised edition. Hobart Press, Summit, New Jersey.
3. Cleveland, William S. 1993. Visualizing Data. Hobart Press, Summit, New Jersey.
4. Robbins, Naomi B. 2005. Creating More Effective Graphs. Wiley, Hoboken, New Jersey.
5. Robbins, Naomi B. 2006. “Dot Plots: A Useful Alternative to Bar Charts,” http://www.b-eye-network.com/view/2468
6. Few, Stephen. 2005. Show Me the Numbers: Designing Tables and Graphs to Enlighten. Analytics Press, Oakland, California, page 59.
About the Author
Naomi B. Robbins is the author of Creating More Effective Graphs, published by John Wiley (2005). She is a consultant, keynote speaker, and seminar leader who specializes in the graphical display of data. She trains employees of corporations and organizations on the effective presentation of data. She also reviews documents and presentations for clients, suggesting improvements or alternative presentations as appropriate. Naomi received her Ph.D. in mathematical statistics from Columbia University. She had a long career at Bell Laboratories before forming NBR, her consulting practice.
I recently picked up Naomi’s book Creating More Effective Graphs. I’m a bookaholic, so it’s surprising that I’ve known of the book for so long without getting a copy. It was worth the wait. It’s an easy read, and because of its simple approaches, clear explanations, and good examples, I’ve rated the book two thumbs up.