thin slice: a thin slice of business intelligence news and views

The last key concept is not new and is almost 100 years old: In 1912 the Gestalt School of Psychology began research into how people perceive patterns, forms and organisations in what we see.  This research culminated in the collection of the Gestalt principals of perception that explain the visual characteristics that cause us to group objects together.

A lot has been written about these, but briefly the characteristics are:
Proximity, Similarity, Enclosure, Connection, Continuity and Closure. 

These all relate to how close, alike, together and aligned objects are that cause our brains to interpret them as grouped objects, which helps us quickly determine aspects about them (whether they are bold or green for example).

The way objects are grouped, aligned or different means we perceive them as groups, this has 2 ideas for report designers:
 - we should group like items together, focus on the value add we are presenting by organising and minimising the data shown.
 - we should separate distinct items, by arranging the information in a way that makes sense, making sure that the important data stands out.

This way we can enable users to quickly interpret the data we are presenting.

But how do we do this?? Thankfully we don't have to work this out for ourselves, there are some very clever and intelligent people who have researched these topics and provided some answers and guidelines.

In the next post, I'll review some people who have been able to help me in my work and help others get the message across simply and clearly. 

Business Intelligence tools provide lots of ways to summarise and aggregate data, however this alone does not mean that the audience will interpret, understand and gain value from it. Only from careful design and planning can report developers and system designers structure the output in a meaningful and value added way.

To enable me to be able to sell others on design ideas and concepts, I needed to understand more about what I saw, how I interpreted data and what I saw as errors.  My research showed me that these were not new concepts and had previously been applied to paper based items not dynamic or interactive screens, which we mostly work with today.

There were 3 concepts which stood out, the first of which is human memory limits:

Memory Limits:

There are 3 types of memory limits, Iconic, Short Term and Long Term.  The important ones for BI developers are iconic and short term.

Iconic memory is very much like a computer memory buffer, where items are held before they are processed - what goes on here is pre-conscious. If we group items (either by size, shape or colour etc) it can help users process the information in iconic memory, called pre-attentive processing.

Short Term memory is the key limitation in human cognition, studies have proved that we can only store 3-9 chunks of visual information at one time in short term memory.  We can help users perceptions here by grouping items and intelligently using charts.

The key for designers is to reduce the short term memory load by using familiar items, objects, actions and directions.

Data Encoding

The key here is how we can visually encode data for faster perception, it's better for pre-attentive processing to occur rather than attentive processing which is sequential and therefore takes longer. The best example is trying to find how many 4s are in the following string:

172634950980273849

It's difficult and slow and there's nothing to distinguish the 4s from the other numbers, in the example below we make them stand out: 

172634950980273849

Much easier!  That's all for this episode, catch the next instalment soon.