Improve Your Dashboard by Understanding Visual Perception
08/31/2019 by Ryan Kumpfmiller Data Communication
Visual perception explains how people ingest diagrams, charts, and dashboards. It is an essential skill for those building reports and data visualizations.
By understanding this trait, your work will become more meaningful to the viewer. You may understand why a bar graph is better than a pie chart, but understanding how folks interpret the visual is essential.
The visual perception process is how the eyes and brain interpret information viewed on a screen. Industry expert Colin Ware has written several books discussing the visual perception concept as it relates to large data sets.
Our visual perceptual process occurs in three quick steps. The process happens so quickly we are often unaware that we are doing it. This visual process is how our eyes and brain work together to form a conclusion about what we are seeing.
Let’s walk through each step as it applies to a dashboard.
In the first step, unconsciously our eyes and brain quickly scan through the dashboard. Here the neurons in the brain pick up the pre-attentive attributes of the charts, such as form, color, position, and motion.
The brain is looking for form constancy. It wants to classify the visual information presented. Is it a text document, is it a dashboard, is it a useful infographic. The viewer is trying to decide if it is worth continuing.
Upon first glance, the brain instantly discerns the presented shapes. It can tell that there are three sections to this report, each unit is colored differently and a different shape.
Maximizing visual learning is about understanding how we perceive images.
The immediate understanding of the webpage is “This is a dashboard that presents three pieces of data.”
After the initial scan, the brain takes a high-level look view of the dashboard. Since the first step broke out the structure, the second step looks at graphics individually to search for patterns and anomalies. This step is where your brain acknowledges oddities in the data visualizations.
The brain’s visual perception is selective.
In the following dashboard, the brain is scanning for differences. Why is one bar so much bigger, why does the bubble chart have a smaller and further away, and what is making that box in the treemap a darker blue?
All of these details are tucked away in the working memory as items requiring more inspection.
Our working memory is limited. Most people can hold three to four elements in their visual memory at once. If distracted, we have problems with visual processing. We struggle to retain the minimum bits of information.
The working memory is holding the few distinct points from step 2 and now looks to add background to what is happening. The brain is seeking a useful takeaway from the data visualization presented.
What are the data elements that are out of place? What measures are being used to make them stick out? Is there a logical reason for this?
These are the questions looking to be answered that long term memory will help build a conclusion around.
The goal of a good visual is to make conclusions evident. You want the user to quickly share your viewpoint.
Memory plays a vital role in data visualization. As noted earlier, the brain’s working memory is limited to a few items at once. When looking at a table of data, this fact means that the brain can only keep a few data fact. Even if they are calculated items in a dataset, the user is still only skimming a fraction of what is available. Your data communication must be clear to be effective.
By creating reports, those few items turn into conclusions that come from consuming large amounts of data graphically displayed.
When creating visual information, keep the visual system in mind. Distinguish general categories of information from each other by color or form, so that they become clear even at the first perceptual stage. Accommodate step two by making key pieces of information stand out from the pattern in some way. Remember that working memory can hold a limited amount of information at once; eliminate all unnecessary information so that you can speed up the analysis process in step three.