James Holiday-Scott

James Holiday-Scott

Gestalt Design in Data Analysis

Gestalt Design in Data Analysis

An introduction and case study analysis.

What are Gestalt Principles?

At the heart of all good graphic design is an understanding and effective application of the Gestalt principles that derive from how we all see and react to what we see.

We all see the world in remarkably similar ways, thanks to many millennia of evolution and a great deal of practice trying to tell the difference between safety and sabre-tooth tigers.

In short, the Gestalt principles are tricks we can all use when we design to make sure we have the mind's attention in the right way.

As data analysts, we want our audience's mind to be calm and focused, and ready to follow us on a journey through our data.

First, we're going to take a brief look at the design principles themselves.

Then, I'm going to use a couple of pieces of data analysis as a case study to explore the Gestalt principles in action.

The Gestalt Principles


We prefer to look at things that are orderly, simple and clear to our eye. Order makes us feel safe and puts us at our ease.

If what we see is not simple, we automatically tend to reduce it to familiar patterns and this is not always a route to accurate interpretation.


Engaging out in-built instinct to find patterns in what we see, closure is when the mind fills in the gaps to complete the whole image. Too little information and nothing happens, but get it right and the mind is subconsciously and actively engaged with our design.

Symmetry and Order

We are all driven to create order from chaos, and therefore have a natural desire for symmetry.

Symmetry is a primary principle - our eyes will find symmetry before proximity.

Uniform Connectedness

The strongest way to tell the eye that two things are related is to connect them in a simple way. A connection will override shape and color and sometimes even proximity.


Again leveraging order, enclosure meets the human need for control. We understand that everything in an enclosure is related, and will conclude everything outside the enclosure is non-related, even when there is no other reason to think this.


Enclosure that is created through space is proximity. Objects can be group by othering them from other objects or groups of objects. Proximity is powerful and can override colour and shape quite easily.


It is our instinct to follow lines to their destination, and then speculate on what happens next. The eye will follow a path made of colour, proximity or connection even when other elements interfere.

Common Fate

Maybe we're all sheep, but we instinctively believe that all objects facing the same way belong together. Elements which convey movement, change and direction can be harnessed to create a common fate.


Combining proximity and common fate, aligning elements in parallel will convey relatedness through pointing, moving, or organising in the same direction.


The mind enjoys matching things - indeed, this is one of the first skills we develop as infants - and therefore the eye is drawn to objects that share characteristics more readily than to objects that do not. The more characteristics in common, the stronger the relatedness.

Focal Points

While many of the other principles engage our desire for calm, order and comfort, focal points work for the opposite reason - they stand out and demand that we identify them as we would a potential danger.

Past Experiences

Symbolism works because we are exposed to white wedding dresses, red stop lights, and yellow fire engines. Only, past experiences vary by culture so this principle is the weakest and least reliable.

Figure and Ground

Think of any optical illusion and you're likely seeing an example of figure and ground. Definitely one for the not-to-do list for data analysts as the effect is confusing and varying visual experiences.

Gestalt Design in Data Analysis

Once all the data-cleaning is done, the bulk of the remaining time will be spend massaging Matplotlib into gorgeous and effective graphic design.

This is a lot easier with Gestalt design in mind, so here are a few examples worked through as case studies to draw out their use of these key design principles as they are applied by data analysts.

The first two examples below come from the beautiful mind of David McCandless. You can see his work at information is beautiful. The last one was created by John Burn-Murdoch, Billy Ehrenberg-Shannon, Aleksandra Wisniewska and Ændrew Rininsland and Steven Bernard for the Financial Times .

1. Proximity, similarity and Symmetry

The purpose of the design is to show that analysis has revealed a stark difference in the mortality rates between men and women.

Essentially a pie-chart, two semi-circles in close proximity communicate that men and women are about 50% of the population. Similarity of shape and proximity does this without a single word, keeping the data:ink ratio high. Symmetry is invited, but denied by the change in size. This creates an unexpected contrast, drawing our eyes inwards to find the statistic of 67%.


Enclosure, Proximity, and Focal Point.

Intended to show the risks of underlying health conditions in COVID-related deaths, the design says a lot with few words.

Enclosure is created by a high contrast cube of cubes in close proximity, intuitively showing the audience what 100% looks like in physical space. A focal point is created by the single blue high-impact cube showing how rare it is to die if you have no underlying conditions. This is the primary message of the analysis.


Continuation and Parallelism

Analysis revealed a startling shift in voting intention - a real data journey through change.

Parallelism is used in two ways. Vertically, the relationship in time is portrayed by the thick coloured bars. Horizontally, parallelism is deconstructed to show how candidates relate - which is by common fate in their movement to Macron. This creates the movement needed for continuation and a common fate. The eye is drawn from top left to top right, then down the right and back left, then up a little on the left before going over to bottom right. We understand the change perfectly with no words at all.


Should Data Analysts Use Gestalt Design?

Yes, definitely. Gestalt design puts the audience first in our minds and breaks down barriers to effective communication. Our job is to make the complex into something intuitive. Our purpose is hide the detail and reveal the truth. Gestalt design helps us achieve both.

Share this