Note: this blog post refers to the book Data Visualisation: A Handbook for Data Driven Design by Andy Kirk.
Let’s begin with a simple question: What is data visualisation?
Defining Data Visualization
I'd like to answer this with a quote from Andy Kirk:
Data Visualisation is “the representation and presentation of data to facilitate understanding.”
Within this definition lie four crucial elements of effective visualization that help convey meaning and insight to the viewer, supporting the primary goal of data visualization: to facilitate understanding.
These four elements are:
- Representation
- Presentation
- Data
- Understanding
These elements are essential for truly understanding a visualization and extracting meaning from it. Let’s take a closer look at each of them.
Data
Looking directly at data, such as numbers in a table, we can examine individual values but struggle to see the bigger picture. When we observe data points individually, a lot of information remains hidden. We need to understand the relationships between them to grasp their meaning. Comparing and contrasting values is the essence of data representation.
Representation
What is Representation in Data Visualization? Representation refers to the decisions made about how data is visually presented. A good representation makes the most effective use of the brain's visual perception capabilities.
Representing data is always a combination of marks and attributes:
- Marks include points, lines, and areas.
- Attributes are the appearance properties of these marks, such as size, color, and position.
These components form the basic anatomy of a chart.
Presentation
The presentation of data goes beyond representation. It considers all other visible design choices—this includes decisions about interactivity, annotation features, and the overall composition of your work.
It is important to note that the above steps are interconnected, not separate. The various steps interact to create the entire design anatomy.
Understanding
Facilitating understanding is essentially the goal of all the steps mentioned above.
The viewer goes through a process of understanding. This involves three stages:
- Perceiving → What does it show?
- Interpreting → What does it mean?
- Comprehending → What does it mean to me?
Let’s explore this with an example.
Imagine a chart showing a football player’s games and goals over a period of 10 years. You might immediately perceive an increase or decrease in the number of goals scored. This is the stage where you simply observe the information.
The next step is to interpret those numbers—this means giving meaning to what you see. At this stage, context matters. For example, if you know football, you’d know that scoring more goals than games played is very impressive and rare. Without this knowledge, the significance of some information might be lost. Here captions or headlines can help guide the viewer’s interpretation.
The final stage is comprehension—drawing personal conclusions about what the data means. What can I learn from this visualization? Does it affect me in any way?
This stage is something the creator usually has little control over, because each viewer is different and will draw different conclusions.
So what constitutes a good data visualisation?
Data visualisation is a medium of communication—not a guaranteed transfer of knowledge, but an opportunity for understanding. The potential for understanding is strengthened by recognizing the key components involved in the process: Data, Representation, Presentation, and Understanding. These elements interact and work together to convey meaning to the viewer.
But what can be achieved when data visualisation effectively incorporates these four components?
Andy Kirk offers a thoughtful answer to this question:
“Visualizers can control the output, but not the outcome. At best, we can expect to have only some influence on it.”
Source: Kirk, Andy (2016): Data Visualisation: A Handbook for Data Driven Design, Sage.