The Gestalt principle of Figure/Ground states that perceive recognisable objects from their background. A Figure lies in foreground and the Ground is the background.

The visual system uses a probabilistic model to determine the Figure and the Ground. Generally it holds that smaller and more recognisable objects are nearly always the Figure and the surrounding with more homogeneous characteristics is the Ground.

The classical example for illustrating the Figure/Ground principle is the Rubin vase, which is a form of ambiguous illusion, and is named after the psychologist Edgar Rubin:

Focusing on the Figure we see the vase, and focusing on the Ground we see two persons facing each other. This visual illusion has an important lesson. Our visual perception works by first detecting the edges of an object and assigning them towards a direction in order to form recognisable shapes. If we assign the edges of this shape inwards we see a vase, but if we assign them outwards we see two persons.

A dominant design feature

The Gestalt principle of Figure/Ground is a dominant design consideration for data graphs. We need to clearly specify which is the Figure and which is the Ground. This is particularly important when designing graphs that deal with multidimensional or complex data.

As a simple example, consider the plotting the Preston curve for the EU member states in 2017, by comparison to the rest of the world. The Preston curve shows how individuals born in richer countries can expect to live longer by comparison to those born in poorer countries. The relation describes a diminishing rate of return to life expectancy, and is of course asymptotic at the top since no one can live for ever.

The graph will show the relation for all countries in 2017, which means that there is considerable overlap of coordinates in a dense scatter plot between economic size (here measured in terms of GDP) and life expectancy at birth (measured in years). Therefore, using retinal variables to differentiate EU member states from the rest of the world, e.g. by using different marker shapes or marker colours, would be an ineffective encoding strategy as it would make it hard to decode any differences.

Instead, given the focus on EU member states (the Figure) compared to the rest of the world (the Ground), it might be best if we show the rest of the world as a homogeneous background that is similar to the axis and other identification information and bring into the foreground only the data related to the EU member states, as follows:

The Figure is the data on EU member states and the Ground is the reference data which is the rest of the world plus the reference identification on the axis labels and grids.

Switching Figure and Ground

Bringing the Figure into focus, as in the example just above, results in the immediate detection of the edges of the foreground objects and the sharp contrast of Figure with Ground.

As an another example, consider the the stacked bar chart just below. The bars represent age groupings and the coloured segments the amount of time that Australians spend connected to social media. The data is sourced form Heads Up, as also acknowledged in the note to the graph.

The white space that is placed in between the bars sharpens the contrast between the Figure and the Ground. We detect the bars’ edges and consider them as contrasting objects. That is, the design of the above stacked bar chart suggests that the graph objective of this data graph is the contrast between age categories. The design assists in locating an age group and learning how much time each age group spends connected on social media.

However, the graph’s title suggests that the graph objective is on ‘How much time Australians spend connected to social media, by age’. The focus is on Australia as a whole and the ‘by age’ grouping is only a secondary objective, more of a reference type of information.

In this case, it is best to switch the Figure into the Ground, by eliminating the white space between the bars and reducing it to a thin reference line. Now the thin white reference lines have become the Figure, and the coloured bars have turned into the Ground by connecting coloured segments across the age categories.

This design shifts the focus into the graph objective of how much time Australia as a whole is connected in social media. It is now clear that the majority of Australia, regardless of age, connects to social media about once a day. Also it is more likely that Australians will spend less than once a day (the purple areas) than more than once a day (the blue areas).

I generally consider the Gestalt principle of Figure/Ground as the dominant consideration when designing a data graph.

Back to Gestalt theory ⟵ ⟶ Continue to Proximity

Demetris Christodoulou