• Votes for this article no votes for this yet
  • Dashboard Insight Newsletter Sign Up

Data Display vs. Data Visualization
A Criticism of Visualization Criticism - Criticism by Robert Kosara

Monday, July 16, 2012

Gregor Aisch recently wrote a posting about gauges, and how he finds them inspiring and beautiful in their simplicity, even though they are generally disliked in visualization. His posting highlights a common misconception about visualization, and a conflation of different uses of data display, that is worth exploring.

Gregor takes issue with the notion that visualization requires a certain number of data points to be displayed. He also considers “breaking those rules” by showing just a single data point in a chart.

Data Display

Let’s take a step back and look at what a speed gauge in a car does, how it works, and how it is used. A gauge consists of a scale with tick marks and numbers, and a pointer needle that points to a value on that scale. When the value changes, perhaps because the car accelerates, the change is reflected immediately. Pointing to a value rather than showing it directly as a number has a lot of advantages. For one, you can use the same mechanical design to show speed, RPM, temperature, etc., simply by wiring it up differently and using a different scale. You can also read the value very easily, especially if you have a scale that has markings to indicate critical values (such as maximum speeds). Some cars have direct numerical displays for speed, and those are much harder to read. In fact, all cars I’ve seen so far with such a display also have some kind of bar display in addition that also shows the speed (though with much less precision). The reason is that it takes a lot more work to read numbers than to watch a needle point, especially when it changes. Gauging acceleration or deceleration is incredibly difficult with a quickly changing numerical display, but very straight-forward with a gauge.

Tasks

The reason gauges work well in cars is because driving is all about current conditions. What you want to know is, how fast am I going right now? How much gas do I have left? What your speed was five minutes ago, or how much gas you had in your tank three hours ago, matters little. The task to be accomplished in driving is simple: change or maintain your speed. That might involve use of the accelerator, brake, gear shift, or whatever, but the initial decision is a simple one: am I going too fast, too slow, or am I at about the right speed? That’s it. Current information, quick action, immediate feedback. A gauge shows one, current, value. That’s what it is good at. Showing more would be possible, of course, but also irrelevant and distracting. In fact, boiling down the amount of information to just what you need is one of the great achievements of car dashboards. There are hundreds of sensors in a modern car, many of which measure values hundreds or thousands of times each second. There is no point in even attempting to show the driver all that information. The only time when the car should show more is when things go wrong. As long as everything is working properly, the driver only needs a small number of current values.

Data Visualization

The case is very different in data visualization. You can split hairs over how many data points you need, but the difference is a qualitative one: visualization shows a lot more data, usually including a lot of history (if there is a time axis), and sometimes even the future (i.e., forecasts). The tasks that a visualization serves are very different, because they typically are much more complex and not simple comparisons of whether a value is too hight or too low. Analysis often requires finding more data, trying out many different views, etc. None of these tasks are comparable to driving a car.

For the rest of the article, click here.

Credit: Eagereyes

Tweet article    Stumble article    Digg article    Buzz article    Delicious bookmark      Dashboard Insight RSS Feed
 
Other articles by this author

Discussion:

No comments have been posted yet.

Site Map | Contribute | Privacy Policy | Contact Us | Dashboard Insight © 2017