It has been a while since I have discussed some of the latest creative thoughts on data visualization from Stephen Few. I have read all of Steve’s books, attended several classes from him, and religiously follow his blog and newsletter on his website, Perceptual Edge.
For those of you who don’t know, Stephen Few is the Founder & Principal of Perceptual Edge. Perceptual Edge, founded in 2003, is a consultancy that was established to help organizations learn to design simple information displays for effective analysis and communication.
Steve has stated that his company will probably always be a company of one or two people, which is the perfect size for him. With 25 years of experience as an innovator, consultant, and educator in the fields of business intelligence and information design, he is now considered the leading expert in data visualization for data sense-making and communication.
Steve writes a quarterly Visual Business Intelligence Newsletter, speaks and teaches internationally, and provides design consulting. In 2004, he wrote the first comprehensive and practical guide to business graphics entitled Show Me the Numbers, now in its second edition. In 2006, he wrote the first and only guide to the visual design of dashboards, entitled Information Dashboard Design, also now in its second edition. In 2009, he wrote the first introduction for non-statisticians to visual data analysis, entitled Now You See It.
Here is his latest thoughts from his newsletter.
Why Do We Visualize Quantitative Data?
Per Stephen Few, we visualize quantitative data to perform three fundamental tasks in an effort to achieve three essential goals:
These three tasks are so fundamental to data visualization, Steve used them to define the term, as follows:
Data visualization is the use of visual representations to explore, make sense of, and communicate data.
Steve poses the question of why is it that we must sometimes use graphical displays to perform these tasks rather than other forms of representation? Why not always express values as numbers in tables? Why express them visually rather than audibly?
Essentially, there is only one good reason to express quantitative data visually: some features of quantitative data can be best perceived and understood, and some quantitative tasks can be best performed, when values are displayed graphically. This is so because of the ways our brains work. Vision is by far our dominant sense. We have evolved to perform many data sensing and processing tasks visually. This has been so since the days of our earliest ancestors who survived and learned to thrive on the African savannah. What visual perception evolved to do especially well, it can do faster and better than the conscious thinking parts of our brains. Data exploration, sensemaking, and communication should always involve an intimate collaboration between seeing and thinking (i.e., visual thinking).
Despite this essential reason for visualizing data, people often do it for reasons that are misguided. Steve dispels a few common myths about data visualization.
Myth #1: We visualize data because some people are visual learners.
While it is true that some people have greater visual thinking abilities than others and that some people have a greater interest in images than others, all people with normal perceptual abilities are predominantly visual. Everyone benefits from data visualization, whether they consider themselves visual learners or not, including those who prefer numbers.
Myth #2: We visualize data for people who have difficulty understanding numbers.
While it is true that some people are more comfortable with quantitative concepts and mathematics than others, even the brightest mathematicians benefit from seeing quantitative information displayed visually. Data visualization is not a dumbed-down expression of quantitative concepts.
Myth #3: We visualize data to grab people’s attention with eye-catching but inevitably less informative displays.
Visualizations don’t need to be dumbed down to be engaging. It isn’t necessary to sacrifice content in lieu of appearance. Data can always be displayed in ways that are optimally informative, pleasing to the eye, and engaging. To engage with a data display without being well-informed of something useful is a waste.
Myth #4: The best data visualizers are those who have been trained in graphic arts.
While training in graphic arts can be useful, it is much more important to understand the data and be trained in visual thinking and communication. Graphic arts training that focuses on marketing (i.e., persuading people to buy or do something through manipulation) and artistry rather than communication can actually get in the way of effective data visualization.
Myth #5: Graphics provide the best means of telling stories contained in data.
While it is true that graphics are often useful and sometimes even essential for data-based storytelling, it isn’t storytelling itself that demands graphics. Much of storytelling is best expressed in words and numbers rather than images. Graphics are useful for storytelling because some features of data are best understood by our brains when they’re presented visually.
We visualize data because the human brain can perceive particular quantitative features and perform particular quantitative tasks most effectively when the data is expressed graphically. Visual data processing provides optimal support for the following:
1. Seeing the big picture
Graphs reveal the big picture: an overview of a data set. An overview summarizes the data’s essential characteristics, from which we can discern what’s routine vs. exceptional.
The series of three bar graphs below provides an overview of the opinions that 15 countries had about America in 2004, not long after the events of 9/11 and the military campaigns that followed.
Steve first discovered this information in the following form on the website of PBS:
Based on this table of numbers, he had to read each value one at a time and, because working memory is limited to three or four simultaneous chunks of information at a time, he couldn’t use this display to construct and hold an overview of these countries’ opinions in his head. To solve this problem, he redisplayed this information as the three bar graphs shown above, which provided the overview that he wanted. Steve was able to use it to quickly get a sense of these countries’ opinions overall and in comparison to one another.
Bonus: Here is a link to where Steve discusses the example above on his website.
2. Easily and rapidly comparing values
Try to quickly compare the magnitudes of values using a table of numbers, such as the one shown above. You can’t, because numbers must be read one at a time and only two numbers can be compared at a time. Graphs, however, such as the bar graphs above, make it possible to see all of the values at once and to easily and rapidly compare them.
3. Seeing patterns among values
Many quantitative messages are revealed in patterns formed by sets of values. These patterns describe the nature of change through time, how values are distributed, and correlations, to name a few.
Try to construct the pattern of monthly change in either domestic or international sales for the entire year using the table below.
Difficult, isn’t it? The line graph below, however, presents the patterns of change in a way that can be perceived immediately, without conscious effort.
You can thank processes that take place in your visual cortex for this. The visual cortex perceives patterns and then the conscious thinking parts of our brains make sense of them.
4. Comparing patterns
Visual representations of patterns are easy to compare. Not only can the independent patterns of domestic and international sales be easily perceived by viewing the graph above, but they can also be compared to one another to determine how they are similar and different.
These four quantitative features and activities require visual displays. This is why we visualize quantitative data.
I attended an online webinar today hosted by Data Science Central titled Making Flow Happen: Dashboards that Persuade, Inform, and Engage. The presenter was Jeff Pettiross (photo, right) from Tableau Software. I found Jeff’s presentation to be very informative and helpful, but it was the Q&A session afterwards that I thought brought an interesting topic to the surface.
The question asked was:
When creating a dataviz and taking feedback, how do you determine what feedback is based on personal opinion and what feedback adds flow to your dataviz?
Jeff discussed this as having principal-centered arguments versus personal-centered arguments. So, for principal-centered arguments, you could refer to Edward Tufte when you are discussing the field of data visualization, junk charts or small multiples, Stephen Few for best practices for dashboard design, or Alberto Cairo for best practices for creating infographics. You could also discuss articles and academic research related to data visualization.
Where the water gets murky is when you are exposed to personal-centered arguments or, basically, someone’s personal opinion. Sometimes when you are sitting in a dataviz review session, the criticism or critiques you receive can feel very personal. Some of it may be in the way the person is expressing their opinion and the intonation in their voice. Other times it truly may be personal; that personal may not like the person being reviewed or feels threatened by their work.
Jeff made a real good suggestion related to personal critiques by simply asking more questions. Deflect the criticism and ask them to tell you more about what they did not like about the visualization. For example, they might feel your dashboard is too crowded or too busy. You might want to ask for suggestions from that person. If the situation allows, you could bring up a copy of that visualization and make the changes in real-time as they are stating their suggestions.
Jeff pointed out that, unfortunately, this will not work in all cases. If you are a paid consultant at a company, and the client insists that they want it a particular way, the old motto “The Customer is Always Right” would take precedence here. You could say, “O.K., we will do it this way this time, but I would like you to consider this as an alternative for future visualizations.”
Jeff pointed out that at Tableau, they are a critique-centric culture. They often have review sessions of their visualizations where people from different areas of the company may sit in. For example, you might have Sales people, consultants, marketing, training, etc. Using thoughtful critiques, spending about 20 minutes on each feature, and including a diverse group of people, they are able to refine the dataviz as a group and learn and hear other people’s ideas on dataviz.
Thanks to Jeff and Data Science Central for a great session today. What do you think? What do you feel is the best way to critique data visualizations?
I would love to hear your thoughts.
Note: I was supposed to have attended Stephen Few’s three data visualization classes this week in Portland, but preparations for a client presentation in Paris next week made attending impossible.
Dashboard Insight posted a synopsis of a post Stephen did on Data Visualization and the Blind. My wife has worked in Special Education for over 39 years and I feel this is a very important topic as we become more of a visual society. I am including the synopsis from Dashboard Insight which also has a link to the full post on Stephen’s Perceptual Edge Website.
One “6 degrees” side note. Stephen mentions in his article that he received an e-mail from Mark Ostroff which motivated his blog on this topic. I knew Mark many years ago. Mark was (and probably still is) the definitive guru on all things Hyperion. I actually went to the Hyperion World Conference one year in Chicago and attended several of Mark’s sessions. A very innovative and passionate speaker.
I will be doing some Paris/France themed data visualization blogs over the next week or so. So stay tuned and laissez le bon temps rouler.
Dashboard Insight had posted an article on data presentation for the blind a couple of weeks ago. There was little information out there on how to handle it and what measures there are for translating information displayed within data visualization or a dashboard to the blind. When Stephen Few posted on this subject a couple of days ago we were quite excited.
“I’ll begin by stating my fundamental position: a dashboard that is accessible to the blind is a contradiction in terms. “A dashboard is a visual display of the most important information needed to achieve one or more objectives, consolidated and arranged on a single screen so the information can be monitored at a glance” (Few, 2005). No forms of data visualization, not just dashboards jam-packed with graphics, can be made fully accessible to someone who is blind. I am not insensitive to the needs of people who are visually or otherwise impaired. I am merely pointing out what anyone who understands data visualization knows: no channel of perception other than vision can fully duplicate the contents of graphs. Similarly, what someone can communicate through the audio channel in the form of music cannot be fully expressed visually. If it could, why bother performing or recording music? Why not just distribute the written score? Vision is unique in its abilities to inform and enable thinking. Those who lack vision can develop their other senses to compensate to an amazing degree, but never in a way that fully duplicates the visual experience.”
“The information that is displayed in a dashboard can and should be presented to people who are blind in a different form when needed… Unfortunately, an alternative form of presentation will not convey all of the information contained in a well-designed dashboard and it won’t communicate the information as efficiently, but if someone who is blind needs the information, it behooves us to provide a reasonable, even if imperfect, alternative. The alternative, however, will not be a dashboard. By definition, a dashboard is a visual display, because the visual channel provides the richest and most efficient means of presenting information for monitoring purposes, which no other channel can match—not even close. If airlines were required by law to provide flight-phobic customers with an earthbound form of transportation, that alternative would be called a train or a bus, not an airplane. In like manner, a means of monitoring that uses braille or a screen reader as its medium should not be called a dashboard. There’s enough confusion about the term already. Let’s not muddy it further.”
It is an insightful and educational read we recommend. You can find the full article here.
The first person who exposed me to best practices in data visualization was Stephen Few. I had the good fortune to take a one day Data Visualization class from him in 2007 at TDWI in San Diego. Stephen’s company is called Perceptual Edge.
Stephen founded Perceptual Edge as a consultancy that was established to help organizations learn to design simple information displays for effective analysis and communication. With 25 years of experience as an innovator, consultant, and educator in the fields of business intelligence and information design, Stephen is now a leading expert in data visualization for data sense-making and communication.
He writes a quarterly Visual Business Intelligence Newsletter, speaks and teaches internationally, and provides design consulting. In 2004, he wrote the first comprehensive and practical guide to business graphics entitled Show Me the Numbers, in 2006, he wrote the first and only guide to the visual design of dashboards, entitled Information Dashboard Design, and in 2009 he wrote the first introduction for non-statisticians to visual data analysis, entitled Now You See It.
With that introduction to Stephen Few, I wanted to provide you a link to his web site and his review of Tableau 8. Here is a brief snippet of Stephen’s review.
“I’ve seen it happen many times, but it never ceases to sadden me. An organization starts off with a clear vision and an impervious commitment to excellence, but as it grows, the vision blurs and excellence gets diluted through a series of compromises. Software companies are often founded by a few people with a great idea, and their beginnings are magical. They shine as beacons, lighting the way, but as they grow, what was once clear becomes clouded, what was once firm becomes flaccid, and what was once promising becomes just one more example of business as usual.”
I have now completed my MOOC course, Introduction to Infographics and Data Visualization, taught by Alberto Cairo. I wanted to write a review so future students of this class know what to expect. I tried to break down my review by topic.
I first found out about the course by visiting Mr. Cairo’s Website, http://thefunctionalart.com, after I had purchased his book to read. I had tried to sign up for the first session of this course taught last year, but it filled up very quickly. I went on full alert to make sure I was able to sign up for the second offering which started last January. Even with a cap of 5,000 students, the class filled quickly, but I was quick and able to enroll.
Mr. Cairo started each week by sending us an e-mail “New message from Alberto Cairo” which had a few notes and a link to the course News and
Announcements forum. In the forum, Mr. Cairo posted detailed instructions for the week along with any recommendations and insights into the assignment. Between Mr. Cairo and Rachel Barrera, his Graduate Assistant for the class, I received e-mails every few days to let us know what the expectations were, informational items, etc. I felt the communication level was just right and both of them answered e-mail questions in a very timely manner.
The lectures were all taught from video. The MOOC philosophy is to keep lectures around 12 minutes or less in length, which works out to about five videos to watch per hour lecture. The reasoning behind this is that our attention span starts to lapse after 15 minutes, so if the class is broken down into smaller chunks, we are more inclined to watch a shorter session on a particular topic as well as retain the information better. For the first week of class, Mr. Cairo’s videos were 2:32 minutes, 6:17 minutes, 12:03 minutes, 8:04 minutes, 9:51 minutes, 14:20 minutes, and 5:35 minutes. His style of lecture is to tell you a story related to the topic. I found the individual lectures very informative, interesting and the time went by very quickly when watching them.
Mr. Cairo gave us a lot of different materials for reading. For example, in the second week of the course, we were assigned the following:
1. Read the interviews with John Grimwade (Condé Nast Traveler) and Steve Duenes/Xaquín GV (The New York Times).
2. Read Data Visualization for Human Perception, by Stephen Few.
Also, each week, Mr. Cairo would provide us links to additional articles, videos, and blogs he put together. They were optional, but again very useful. He also sent us an e-mail each week of links to other interesting materials to read.
Each week, we were required to participate in the discussion forums. Whether it was to post our opinion on a topic or review other classmates assignments, we had to post 2-3 entries each week. At first, I did not think I would like this, but found this to be one of the most enjoyable parts of the class. When reviewing other classmate’s projects and assignments, we had 5,000 different examples to choose from, so there should have been discussions that appealed to everyone. I was very fortunate since the ones I picked were very interesting to read. Since we had a large pool of people from many different walks of life, we had a lot of diversity in why they created the design they did, their personal or professional interest in that topic, and the actual visualization they produced often gave me ideas for projects I was working on at work. Even after I finished my mandatory 2-3 entries to review, I often went back and read others I thought were of interest. For the final assignment we were able to pick our own topic. I frequented the discussion forum a lot just to see the variety of topics and infographics my classmates created. I was a bit frustrated that time did not permit me to view them all.
We had two quizzes early in the class. If you read the materials and watched the lectures, you will have no problem with these.
We had three projects to complete as part of the class. The first was to create a topical interactive graphic. The second was to create a visualization, and the third project was to create an infographic.
I put a lot of time into these projects. I was fortunate in that I was able to tie my third project into a need we had at work for an infographic. So, not only was I learning, but I was able to promote the use of infographics at work.
For the second assignment, I really liked the visualization created by one of my classmates I will refer to as “Jim.” I liked it so much in fact, that I wanted to make a working example for our development team at work. I create a lot of dashboard “templates” for our development team in MicroStrategy, which is our enterprise standard BI tool.
So, using Jim’s data and format, I created a dashboard in MicroStrategy with some tweaks to it.
I have included a screenshot of my assignment on this project below.
I used horizontal stacked bar charts instead so that the viewer can visually see how social security and income tax rate add up to the total and explains visually why the countries are ordered the way they are on the dashboard. I also separated out $100K and $300K percentages into separate visuals.
In addition, I added the flags of the countries. Yes, I know, chart junk!
Now, you don’t see any numbers on the data points in this dashboard. The reason you don’t see them is because they appear when you mouse over a bar where you then see the country, category and the percent value as a tooltip.
I don’t know Jim but want to thank him for providing a great example for me to follow. This will help our team a lot in creating future dashboards.Best Quote
“Christmas cards do not cause Christmas to happen, but the two are highly correlated in time.”
I know there is a lot of discussion about MOOCs and if they adequately provide a viable learning device right now. I feel a MOOC is really no different from any University course I took when I was in college. You will get out of a MOOC course what you are willing to put into it. I took this course very seriously and set my getting the Certificate of Completion as my goal. To get this, I had to do all of the course work. This class was something I wanted to take to enhance my skills as well as my career. I also took this course because I had read Mr. Cairo’s book, The Functional Art, and wanted to learn more from him. In regards to the readings, I was fortunate that I had already read most of Mr. Cairo’s book and had previously read many of the articles he assigned, such as Stephen Few’s material, so the reading assignments were not as steeped for me. However, I did go out and read a lot of the supplemental materials that I found of interest too.
The lectures were excellent and some were down right fascinating. I loved how Mr. Cairo told the story about John Snow and the 1854 Cholera Epidemic in London. I also loved the story and explanation of how we interpret circles and why not to use them in data visualizations.
When Mr. Cairo offers his third version of this class, I highly recommend you take it if you have the opportunity (sign up early!) I find myself longing for more and hope Mr. Cairo or his counterparts like Stephen Few, Nigel Holmes, Colin Ware or Edward Tufte offer similar MOOC courses.
Before I discuss this graph, I wanted to point out that one of my co-workers sent this to me. I send a lot of data visualization examples around the company to expose our business partners to different ways data can be seen. This is important because too often during a requirements session, someone will ask the business partner if they need a dashboard. This is problematic on two fronts: first, many are not sure what a dashboard is and how it will benefit them, and second, even if they know what a dashboard is, they are not sure what makes up a dashboard (e.g., graphs, selectors, etc.) It’s a case of “How do you know what you don’t know.”
My mantra is that I want to get my business partners excited about their data (sorry, Dr. Phil for stealing your tagline). The fact that after a year of me providing data visualization training, dashboard design best practcies, etc., the business is now getting excited to the point where they are sending me cool visuals that they find. I am exicted by the fact that they are doing this and take a little pride that we have built this collaborative, sharing relationship.
Now on to this graph (Thanks Rich!).
In 1980, the five most commonly spoken languages other than English were Spanish, Italian, German, French, and Polish. By 2010, Spanish was still the most widely spoken language after English but it was followed by Chinese, French, Tagalog, and Vietnamese. [SOURCE]
NOTE: Square area is proportional to the number of people speaking a given language. Data are for the population 5 and older.
Spanish includes Spanish Creole, French includes Patois, Cajun, and French Creole; Portuguese includes Portuguese Creole. The languages highlighted are the languages where comparable data were available for the four time periods: 1980, 1990, 2000, and 2010. Other languages spoken in 2010 at least as widely as those shown above include Arabic, Cambodian, Gujarati, Hebrew, Hindi, Hmong, Navajo, Thai, and Urdu.
As is the case with all surveys (including the 1980, 1990, and Census 2000 “long-form”), statistics from sample surveys are subject to sampling and nonsampling error.
This is definitely a very interesting and engaging combo chart. Stephen Few and Alberto Cairo caution against using circles since a reader will tend to want to compare the diameters of multiple circles versus the correct way of wanting to compare the areas of the circles. I am curious what they would think of these squares and the legend provided to interpret the squares. I will revisit this circle topic in a future blog.
On Tuesday evening, January 29th, 2013, Apollo Group was bestowed the Excellence in Dashboard Design Award at MicroStrategy World 2013. Their entry, a Student Performance Dashboard, was based on portions of the top three entries in Stephen Few’s Dashboard Design Contest that was held last year. Unlike those entries in Stephen’s contest, which were designed in Photoshop and Excel, Apollo Group’s entry was design using MicroStrategy v9.2.1 and the Visualization SDK (Adobe Flash Builder Professional v4.0.1/Flex SDK v4.1). Apollo Group’s entry was one of the top 5 entries (no order given) based on 89 submissions; There were also several other entries that received Honorable Mention.
My entry for the MicroStrategy World 2013 Dashboard Contest is a prototype of a Student Performance Dashboard. This dashboard will be part of a new suite of higher education products and services that collect and organize key operational and performance data that deliver actionable metrics and analytics. This new suite of products is known as Apollo’s Education-as-a-Service (EaaS) and is referred to as AES.
The actual implementation of this dashboard would be used by faculty to show all of the student performance data, for a specific class, on a single screen with the goal for the instructor to immediately visually understand the key performance metrics and take action on them.
Back in August of this year, Stephen Few, data visualization evangelist and author of the seminal book, Information Dashboard Design, announced a contest to design a dashboard following best practices and principles. The contest required participants to design the dashboard using student performance and assessment data that Stephen provided. Any graphic design tool (e.g., Photoshop, InDesign and Excel) or BI tool could be used to create the dashboard.
The winners were announced in October of this year. There were 91 entries. The contest focused more on innovative dashboard design principles rather than the use of BI tools. The winners and the tool they used are:
1st Place: Jason Lockwood Photoshop
2nd Place: Shamik Sharma Excel 2010
3rd Place: Joey Cherdarchuk Excel 2010
To the best of my knowledge (and Stephen’s), none of the participants used MicroStrategy to create their dashboard. A few of the participants did use Tableau and SAS. This fact alone made me want to create an innovative dashboard to demonstrate the capabilities of MicroStrategy (Disclaimer: I am not an employee of MicroStrategy and chose to use this tool since it is our internal standard BI tool. I am not endorsing MicroStrategy or any other tool for the purpose of creating this dashboard).
Each of the three winning designs contained elements the Apollo Group was interested in for our proposed Student Performance Dashboard for EaaS. I wanted to be able to incorporate elements from each of these winning dashboards into a single dashboard so that I had a prototype to show our internal business partners. In addition, I wanted to be able to demonstrate to them the capabilities of MicroStrategy.
Stephen Few is using the examples from his contest in the second edition of his book, Information Dashboard Design, which should be out later this year. I have been a big fan of Mr. Few’s for many years and encourage you to purchase this book once it is published to see and read more in-depth insights on the dashboards created for his contest. I also highly encourage you to visit his blog at http://www.perceptualedge.com/.
Below is a screenshot of our award-winning MicroStrategy version of this dashboard. I have also attached a PDF file of my summary overview of the dashboard that I submitted as part of my entry to MicroStrategy.