Human cognition has its limitations encompassing
- producing and understanding language
Good working dashboards accommodate those limitations of cognition and simplify the user’s job, which includes
from the body of information represented in the visual display.
- identifying trends
- drawing conclusions
One way to support human cognition is by limiting the number of items the user must retain in short-term memory, when navigating and using information in the display.
Another way of supporting human cognition is by reducing the formatting that users have to retain in short term memory, such as
Moving them into a legen besides a chart such as the pie chart, requires users to remember the key, and constantly look up the information again and again. It is better to use a display method, where these can be incorporated into the visualization itself.
As designers we want to avoid pie charts in dashboards, because they typically require a legend next to it in order to be interpreted correctly. Readers are then forced to remember the often color encoded pairs, and recall them when reading and interpreting the pie chart.
We designers could label the segments accordingly, but as with the nature of dashboards, data changes, and hence the labels would become illegible. It’s hence better to use an alternative display for the visualizaiton of data.
Short-term memory is limited in capacity and duration that information can be retained. The capacity is limited to a maximum number of 9 items, whilst the duration to retain the information is limited to minimum 9 to maximum 18 seconds.
Because of these limitations we cannot extract relevant information from visual displays that ask us to perform more. Any interruption that requires cognition to perform a “side” task poses a risk in getting lost in the task the user actually tried to accomplish.
Also an overwhelming amount of data that uses color codes which readers have to retain, are difficult to decipher. Because of this it is best to use a display method in information visualization that shows the meaning of the data represented contextual to the visualization.