Comments on: Is Data Visualization Useful? You’ll Have to Prove It.

By: James Lytle

James Lytle — Wed, 06 Jan 2010 15:13:42 +0000

I think what Stephen is trying to get at here is that you need to have some legitimate grounding for an envisioned sense of “candies.” Poor expressions interpret candies as extra colors and swooshy animation that more often than not simply add distraction from the task at hand. I think your ratio is a bit off too. We need to educate through evidence, and, as those studies grow, as proper evaluation methods are established akin to the field of usability, the formerly uneducated will be disappointed it came so late. Take some time to read legitimate studies in this realm, often referred to as information aesthetics. Andrew Vande Moere from the University of Sydney is one of few slowly making headway (“Towards a Model of Information Aesthetics in Information Visualization” is a good place to start.) Still, though recorded evidence is elusive, we need to at least build a proper mental model and common vocabulary for the aesthetic principles so we can consistently, systematically respond to requests to ’spice things up a bit’.

By: Stephen Few

Stephen Few — Tue, 05 Jan 2010 01:34:47 +0000

Jorge,

Although I appreciate your blog, I believe that in an effort to be provocative you sometimes stray from the facts. You have done so in this particular post.

You wrote: “When Stephen Few asks the readers “true stories about the benefits of data visualization” that’s almost an admission of impotence.” Not at all. The effectiveness of data visualization is well established by a large body of empirical evidence. My request for “true stories about the benefits of data visualization” was an attempt to get people to share their real-world successes with others. People have told me many real-world success stories, but I have never bothered to record them, and therefore can never manage to remember them in detail.

You wrote: “If a single chart can save the world, it will not be a Few’s or Tufte’s 100% compliant chart. It will be a glossy Xcelsius pie chart.” Didn’t you mean to say “It will “not” be a glossy Xcelsius pie chart”? Based on your later statement about pie charts, I suspect that you did. I agree that if a single chart could save the world, it certainly doesn’t need to be 100% compliant with the principles that I teach, but you said that “it will not.” You’re being provocative in a way that could backfire on you and any of us who try to help people use data visualization effectively. There are a great many software vendors out there producing really ineffective products that love any opportunity they can get to promote their bad products. They love statements like yours and will take them out of context to serve their own interests in ways that will harm the people who need good data visualization tools.

You ended with the statement: “Draw a line but don’t forget the candies. You can take a horse to the water, but you can’t make him drink, unless you give him some sugar cubes.” Horses that are thirsty don’t need to be tempted with candy. Pure water running freshly in a stream is what they want. People that aren’t thirsty might need to be tempted with something sweet, but you don’t give them candy instead of water, you give them candy to get them to then drink the water. You seem to be suggesting something about my work that you’ve suggested in the past–that I oppose charts that are delightful in appearance or interactive. You know better than this. The “eye-candy” and interactivity that I oppose is only that which undermines the objectives of the chart. Beauty and enjoyment need not be at odds with effectiveness. People who create charts that incorporate decoration and motion in ways that undermine meaning, distract from the message, or that make the charts difficult to understand are unskilled designers. Do you believe otherwise?

To your reader PragmaticCynic, who commented above, I’d like to say that you must not be familiar with my work. “Ivy-towered pretentiousness of what is good”? You won’t find any examples of this in my work. If you believe otherwise, I invite you to share them. My work seeks to help people like you who present information to those who have “little time to study a chart.” What I teach is extremely simple and practical. It is designed to “hold their attention” and “make it interesting” in ways that engage them with the data. This is something that, contrary to your claim, I never forget. Anyone who knows my work knows that I don’t teach lofty techniques that are designed for “The Royal Economic Society.” You are mistaken.

And finally, Jorge, you not only failed to correct Pragmatic Cynic’s errors, you actually fueled them by saying that my “approach to information visualization partially fails because people’s emotions are removed from the equation.” I do not discount people’s emotions. I am well aware of them and the ways that they can both enhance and undermine the effectiveness of data visualization. If you believe otherwise, as you suggest, then you should support your claim with concrete examples.

By: Michael

Michael — Thu, 29 Oct 2009 12:54:51 +0000

This a solid pragmatic and prosaic piece. The person communicating should not blame those who are receivers … there is no control there. You need to re-evaluate how you are communicating and connect to the audience. I like data and stats and more complex charts – most people don’t. If I lose them, I’ve missed my objective. Simple as that. Know thy audience and what they can comprehend and digest.

By: michaelgalloy.com » Is Data Visualization Useful?

michaelgalloy.com » Is Data Visualization Useful? — Tue, 27 Oct 2009 17:39:21 +0000

[...] Camoes has an article on his Charts blog about whether all the fancy points and rules about visualization are really [...]

By: Dan Murray

Dan Murray — Tue, 27 Oct 2009 12:47:21 +0000

Personally, I’ve received a great deal of value from both Excel, Tableau and the combination of using Excel as a poor-man’s data warehouse in conjunction with Tableau.

I find 3D charts very annoying, always have. I think Few makes very good points about what is good and what is not so good. As this market develops and people have better understand of the differences between simple data visualization and the ability to rapidly probe large data sets, I think people will see the value in tools like Tableau.

Keep up the good work on this blog. I enjoy reading it. Thought provoking and practical.

By: Hugo

Hugo — Tue, 27 Oct 2009 10:08:36 +0000

Just a quick note of hope: since I’ve started reading Jorge’s blog (and others of the kind) I’ve started to implement some of the good advices that I’ve found. Slowly, obviously.

So, what started to be a numerical monthly report, w/ almost no visualization – good or bad – is now almost completely revised. We can actually look at the report and identify some trends…

Sadly, this has taken about a year But, it has made by job easier and it has made it easier to get my proposals accepted by my Manager (who is open minded in what regards this changes).

For a room full of people and/or the board, I still go with flashy charts (awful pie charts, colored, and with no real use). It is, has you have well said, all about getting their attention and make them accept our message…

By: Bernard Lebelle

Bernard Lebelle — Tue, 27 Oct 2009 08:33:16 +0000

Jorge, I totally agree with your pragmatic view on the usefulness of data vizualisation. In my opinion, the current business lack of adoption of dataviz best-practices is mainly due to 2 elements:
* dataviz best-practices not being part of your standard school/graduation curriculum
* user-brain formating based on defaults options in software and exposure to a small set of standard graphics in both press/tv media
Trying to prove the usefulness of a more sophisticated representation is a lot of work because it implies trying to change the mindset of the reader. Most of the time it’s a long process… but for those few situations where you hit spot on with a “best practice” approach that appeals to the leader of the organization.

By: Jorge Camoes

Jorge Camoes — Mon, 26 Oct 2009 23:23:42 +0000

PragmaticCynic: Tufte says “If the statistics are boring, then you’ve got the wrong numbers”. It’s not that simple. His (and Few’s) approach to information visualization partially fails because people’s emotions are removed from the equation.

Nick: I like to think that a good chart, like a good design, is invisible. Unfortunately, we don’t always recognize how important some invisible things are…

Jerome: most people don’t even know what “data visualization” means! On the other hand, they think they know what a chart is. And they know how to make one in Excel. That’s our starting point. Discussing “working memory”, “data-ink ratios” or (God forbid) “using color preattentively” is, not only boring, but also pointless. First stop: “let’s make some simple adjustments to show how they help you to understand your data better”. (To teach someone to ride the data visualization bicycle we need training wheels). Yes, visualization can change the world. One user at a time. At this rate, 2010 is perhaps too optimistic…

By: jerome cukier

jerome cukier — Mon, 26 Oct 2009 22:34:09 +0000

Jorge, I was at visweek this month and was debating a similar topic: can visualization change the world? there are quite a few examples were a visualization was interesting, thought-provoking, etc. but I don’t think that the discipline has lived to its potential, because it’s not used enough. there are tons of people out there, especially in businesses, who could benefit from insights gained from better visual practices. Such techniques are gaining ground, however, because of the availability of new tools and more data. At last year’s visweek, the CEO of tableau gave a keynote speak titled “Practical applications of visual analytics: On the cusp of widespread adoption”. I see this happening in 2010.
I did meet Stephen Few over there and asked him if he had gathered any good stories, and was disappointed that he hadn’t. That doesn’t mean that visualization is not useful, but that, again, it’s not used enough.

By: nick gogerty

nick gogerty — Mon, 26 Oct 2009 19:23:56 +0000

very well stated. The medium (colors and whiz bang) should never distract from the message. And correctly to your point a lot of people are looking for attention, thus showing off with the tools instead of getting the job done.

The best mediums convey meaning without drawing attention to themselves. The economist has excellent chart design that in an english sense doesn’t draw attention to itself, it just performs. That is of course with the exception of the horrific “year in 2010″ and other year in Economist publications. These year end predicotorama feast appear to be the handy work of USA today chart makers on a ritalin binge. Novelty and trite tricks attract attention, consistent quality endures.