Category Archives: Dashboards

Dashboards, Schdashboards

Read an article on the psychology of informational dashboards by Shilpi Choudhury  today. In it, she summarises the fundamental reasons why humans crave for information summarised on dashboards:

[1] Humans have a natural desire to be in control.

[2] Humans have a notoriously short short term memory.

[3] Humans desire to cut down clutter.

My thoughts

I guess being the pragmatist I am, the next question that naturally comes to my mind is then: How do we create appealing, effective dashboards?

Some things came up to mind:

Reducing cognitive clutter

Shilpi cites an article by Jakob Nielsen and mentions that we typically can only remember up to 7 distinct chunks of information in our short term memory, and that this information fades in about 20 seconds. What this suggests to me is that we can’t exactly perform cognitive processes involving more than 7 distinct pieces of information — so excel sheets can actually hurt more than help.

Here’s where presenting data visually becomes important – graphs, info-graphics and the likes all serve to reduce the cognitive load such that we can absorb, understand and memorise (short term) more┬ádata in order to perform more complex analysis and trending, in effect allowing us to see the forest for the trees easily.

[Likewise in communications, people tire when seeing too many pieces of information/data and would much rather welcome an image]

Of course, not all info-graphics were created equally. The well crafted info-graphic utilises natural psychologically distinct visual attributes like colour, size, density, shape (the gestalt principles!) in order to compress information and hence convey more information to the viewer at a quick glance.

Timely Reminders, Proper Organisation

A second point that would make dashboards useful would be the way it can coordinate and inform users when and where intervention is needed. There should be flexibility for intervention rules to be set in order for the dashboard to prove valuable to the user, whose requirements might change across time.

Finally, a proper dashboard calls for proper organisation of data – and this honestly, shouldn’t be done at the level of the dashboard, but rather at the data architecture level. Ultimately a dashboard is meant for the display and manipulation of data that has been captured and organised within a system – if the system has a crappy data architecture, then the dashboard is probably not going to be any bit helpful.