Gauges: The Black Sheep of Data Visualization
Gauges are like the junk food of dashboards. Just as junk food has little nutritional value but contains too many calories, gauges don’t contain enough data for the space they occupy on a dashboard – little “nutritional” value and too many pixels.
That being said, gauges are still a widely used form of data visualization, but they don’t need to be. The wide use of gauges can mainly be attributed to a lack of awareness of the superior alternatives to them. What separates gauges from the rest of the visualization herd? When it comes to dealing with business data, gauges have three main faults:
- Space - Gauges take up far too much of it and can easily be replaced by visualization elements that display the same information in a smaller area.
- Comparisons - Gauges are terrible when you need to compare data, bar or column charts are much more effective at being able to compare numbers against each other.
- Trends - Because gauges don’t display information over time, it is impossible to get the whole picture from the data they represent. When you need to provide historical context to your data (data over time) you typically want to use a line chart.
Gauges usually display a single key measure with the outer scale of the gauge often color-coded to provide additional performance context (green for good, red for bad). Bullet graphs, however, can effectively display the same information as a gauge but occupy much less space. Dashboard should fit onto a single screen. When building dashboards, you want to maintain a high data to ink ratio. This way you ensure that you optimize the space being used.
The graphic below comes from Perceptual Edge and outlines the parts that make up the typical bullet graph.
Compare the gauge and the bullet graph yourself:
The circular gauge takes up more screen space here than the bullet graph, 51070.5 pixels compared to 14430 pixels to be precise. While we could stand to decrease the size of the bullet graph further, the gauge doesn't have much room to play with in this regard.
Now, what if you had two sets of data that needed to be compared? Placing gauges beside each other is much less effective than displaying them in a bar chart.
Now, this one is an obvious choice. With each of the gauges only being able to display one set of data, your eyes have to move from one to the other, taking more time for your brain to process and compare the information. While with the bar graph, multiple sets of data can be used and presented side by side, making it quick and easy to compare the information.
This highlights another downfall of the gauge: it cannot display trends.
Because gauges typically display a single value and range, they are not able to provide us with the whole picture of what might be happening. For example, a gauge may display sales levels for December with the indicator in the acceptable but not good range. With just this data we might assume things are fine, not great but fine, and not give it much more thought.
However, if we were to look at a line chart displaying other months or years we could end up with a number of different interpretations depending on the trend being displaed. If we have typically been in the good range during the other months, or Decembers in past years, then we may need to take action to find out why this is not the case now. Or line chart has provided us with important context that a gauge cannot.
When it comes to business data, it is obvious that gauges are sorely lacking in usefulness, but they aren’t totally obsolete. In our metaphor they may be the black sheep but still sheep nonetheless. They’re still used and are useful in cars, machine and pressure monitoring, and the physical analogue world (a thermometer is a linear gauge for example). But I think we can all agree that when it comes to business and business data, gauges are better left on the digital cutting room floor.