Data + Science

[ad_1]


2/1/2018
The Definition of a Dashboard



What is music? What is art? What is an automobile? There is a time and place for carefully drafted definitions. For example, when drafting legal contracts it is very important to have defined terms so that all parties to the contract have a clear understanding and avoid confusion.

In the opening pages of The Big Book of Dashboards, we defined the term dashboard:

“A dashboard is a visual display of data used to monitor conditions and/or facilitate understanding.”

We spent a good bit of time debating this definition. We all recognize this is a broad definition, but we came to the conclusion that business dashboards, their uses, and designs, are so broad that it’s not possible to create a narrow definition.

Stephen Few recently wrote a blog post “There’s Nothing Mere About Semantics” where he criticizes me and my fellow authors for what he feels is an overly-broad definition.

The purpose of visualizing data is to gain insights. In the past 20+ years, I have never once been in a meeting where someone has asked, “What is a dashboard?” or said, “I really like that faceted analytical display.” These terms are meaningless to people in the business world.

The purpose of visualizing data is to gain insights. In the past 20+ years, I have never once been in a meeting where someone has asked, “What is a dashboard?” or said, “I really like that faceted analytical display.” These terms are meaningless to people in the business world.

Nevertheless, since Mr. Few decided to write his post, I wanted to address some points about this definition.

Let’s examine Few’s definition of a dashboard. In both his 2004 and 2013 editions of Information Dashboard Design, his definition is:

“A dashboard is a visual display of the most important information needed to achieve one or more objectives that has been consolidated on a single computer screen so it can be monitored at a glance.”

My co-authors and I have tremendous difficulty with his definition. My biggest problem with his definition is the second part, “consolidated on a single computer screen so it can be monitored at a glance.”

Why does a dashboard have to be consolidated to a single screen? He claims that this is important because we need to see everything at a glance in order to monitor it. Certainly there are instances where a consolidated view is necessary, but that’s not always the case. If a dashboard is optimized for a mobile device and requires some scrolling, why would that make it any less of a dashboard? Given the advances in technology, we can now swipe and scroll with ease, or navigate with a push of a button. In fact, people expect to scroll on a tablet or smart phone.

Why does it have to be monitored at a glance? We live in a society of instant gratification, so I get that we want answers quickly, but not every question can, or should, be answered at a glance. The fact that a dashboard takes more than a glance shouldn’t invalidate it as a dashboard.

Maybe Few changed his mind on these points. His most recent definition of a dashboard that he uses in his workshops is different:

“A dashboard is a predominantly visual information display that people use to rapidly monitor current conditions that require a timely response to fulfill a specific role.”

Mr. Few uses many qualifiers which lead to a very specific definition. However, some of these qualifiers are, at best, just clutter, and worse, lead to an overly-specific definition creating exception after exception. A dashboard is:

predominately visual – How much is enough or too little?
that people use – as opposed to chimpanzees? This isn’t needed.
rapidly monitor – this changed from “at a glance” in the previous definition. How does “at a glance” differ from “rapidly”?
current conditions – what is the definition of current?
timely response – what is the definition of timely?
specific role – what is specific about the role?

Why stop there? If this were a legal contract, each qualifier would also need to be a defined term. Or we can write it out in detail.

“A dashboard is a predominantly visual information display, shown on a computer screen, that people who are predominately 18 to 65 years old, use for around 1 minute, to monitor conditions, that are often immediate but no more than 3 months past, because the monitored condition requires a response that is typically between immediate and 24 hours, but no more than one week, to fulfill a specific role that must be defined by a business objective.”

While Mr. Few strives for clarity in his definitions, he contradicts himself. He states in his most recent blog post that we as authors of The Big Book of Dashboards have “added to the confusion that exists about dashboards and dashboard design”, but let’s examine how Few himself has contributed to this confusion over the years.

In the first edition of Information Dashboard Design, Few categorizes dashboards into three roles; Strategic, Analytical and Operational (page 40-41).

Here’s what he wrote about Strategic dashboards: “Given the long-term strategic direction, rather than immediate reactions to fast-paced changes, these dashboards don’t require real-time data.”

His second category of dashboards are Analytical dashboards, or dashboards that support data analysis. Few writes, “Analytical dashboards should support interactions with the data, such as drilling down into the underlying details, to enable the exploration needed to make sense of it.”

This is not the only category that Few defined. He outlined a table of variables and values to structure a taxonomy of dashboards. In addition to the three roles above, he also lists quantitative vs. non-quantitative, static display vs. interactive display, primarily graphical vs. primarily text, portal functionality vs. no portal functionality and balanced scorecard (for example KPIs) vs. non-performance.

Contrast this to the second edition of Information Dashboard Design below (page 30).

A dashboard is NOT any of the following: [emphasis added]

• A display that is used for data exploration and analysis
  [Once its own category of dashboards as defined by Few is no longer considered a dashboard.]
• A portal
• A scorecard
• A report that people use to lookup specific facts

These contradictions aren’t simply a result of a new edition of his book. In the second edition (page 226) he features a Marketing Analysis Dashboard, serving much of the same purpose as the Web Analytics dashboard in Chapter 13 of The Big Book of Dashboards – monitoring things like visitors and pageviews. Few writes, “Like all dashboards, it is used to monitor the information needed to do a job, but in this case that job happens to primarily involve analysis.” [emphasis added] Many of the dashboards in The Big Book of Dashboards do exactly that, monitor information that involves analysis.

Another example is the final dashboard shown in Information Dashboard Design (2013, 2nd edition). Cited as an example of a good dashboard, he features a “World Population Dashboard” designed by Jorge Camoes (page 232-233). In describing it, Few writes, “this is not a true performance monitoring dashboard…but is a variation that is designed for the general public to access and view on the web.”

Few himself, so devoted to his definition, included this World Population dashboard in his second edition of Information Dashboard Design in 2013. How does this example fit his definition “rapid monitoring of current conditions that require a timely response to fulfill a specific role”? What’s the timely response for the reader? Go have more kids? If it’s not a “true performance monitoring dashboard” then what is it?

Do you see the problem? Few labels this as a dashboard in his own book, but then calls out others as “adding confusion.”
This World Population Dashboard serves the same purpose as a number of the dashboards in our book, for example, the Big Mac Index designed by The Economist, or The Financial Times dashboard showing the economy at a glance.

Few ends his blog post by exclaiming that in The Big Book of Dashboards “only 2 of the 28 examples of displays that appear in the book qualify as rapid-monitoring displays”. Never mind the fact that we don’t use the word “rapid”, we believe that the majority of the dashboards in The Big Book of Dashboards are in fact designed to monitor conditions to achieve an objective. Furthermore, even if they don’t, there is certainly a related scenario listed in most of the chapters where the design elements could be used to do so. This is an important distinction, because that was the whole purpose of the book.

Few’s Information Dashboard Design has over 50 poorly designed dashboards and detailed explanations about why they don’t work. Open the second edition and see how far you get into the book before you find an example of a good dashboard. I think you’ll discover the first one on page 153, and it’s the dashboard featured on the front cover (also, very similar to the design and function of the dashboard in Chapter 10 of The Big Book of Dashboards). In contrast to his book, we tried to find examples of good design and explain why we think they work for their intended purpose. This is not to say that presenting bad examples with commentary isn’t helpful in the learning process, but we wanted our book to have more examples to emulate. We chose the 28 scenarios to demonstrate real-world business problems and offer potential solutions for them, discussing why we think they work (and we also include what we might improve). It’s our hope that our readers will take inspiration from our book when they encounter similar situations in their work.

Back in 2007, a well-known author in the field of data visualization wrote:

“I want to focus on the insights that emerge when software enables us to view a set of data from several perspectives at the same time. I call a screen that contains multiple concurrent views of a common data set so that comparisons can be made a ‘faceted analytical display.’ Don’t worry about the name, however. Whatever you choose to call it is fine. What
matters is that you have the means to expand your analytical reach by viewing data in this way.
[emphasis added]

Who wrote these wise words? Well, this was actually written by Stephen Few in his whitepaper entitled “Three Blind Men”. It’s worth repeating those last two sentences. “Dont’ worry about the name, however. Whatever you choose to call it is fine.”

Ok, I choose to call these displays Dashboards. As do many other data analysts working in the field every day. As do the major software vendors enabling an unprecedented explosion of data analytics in the real-world on a daily basis. We used the term “dashboard” because that’s what the world of data analytics calls them, and we wanted to help those people make their dashboards better.

Later in that same article, Few writes again, “Here’s an example that I quickly threw together using the new multi-chart functionality that Tableau calls a dashboard, but I prefer to call a faceted analytical display to distinguish a display like this that is used for interactive data explorations and analysis from one that is used to monitor what’s going on.”

This is another example of confusion and contradiction. In this 2007 whitepaper, he “prefers” to use the term faceted analytical display for “interactive data exploration and analysis”, but just 4 years prior to that, he defined this same thing as an “Analytical Dashboard” in his dashboard categorization and taxonomy (Information Dashboard Design, first edition, pages 40-41).

If our goal is to help people visualize their data in a meaningful way, then I don’t think our time and effort are best spent arguing for a tightly-scoped definition of a widely-used term. This provides no value in helping people visualize their data in a meaningful way. In the opening sentence of The Big Book of Dashboards, we acknowledge Stephen Few, “whose books have made a profound and lasting impression on us.” We were inspired by and learned from these books, despite the fact that we disagreed with his definition of a dashboard or found contradictions in his writing, or that we wished for more good examples of dashboards. Few’s first three books helped pave the way for us. Unfortunately, instead of continuing this great work, he has decided instead to focus on policing terms like “big data” and “data science”, and constantly criticizing marketing departments and business intelligence software vendors.

“If our goal is to help people visualize their data in a meaningful way, then I don’t think our time and effort are best spent arguing for a tightly-scoped definition of a widely-used term. This provides no value in helping people visualize their data in a meaningful way.”

It’s hard to find the signal in all of the noise. This debate is the noise, and it’s serving to distract from the signal: how to visualize data in a way that helps all of us understand the world we live in. In the end, that’s what we learned, in part from Stephen Few, and that’s what we hope to do with our book.

Jeffrey A. Shaffer
Follow on Twitter @HighVizAbility



Read More …

[ad_2]


Write a comment