datamake

Towards some principles of visual poetics · final | Arrival

This series humbly looks up to Aristotle asking him how his principles of storytelling (plot, character, thought, diction, song and spectacle) can be applied to, and potentially enhance data visualisation. After an introduction we’ve covered Aristotle's principles of storytelling plot, character, thought, diction, song, spectacle as well as the story-aim: catharsis. This last post wraps our mission up with some summarising fireworks to take home…

The visual principles

Now then. What has Aristotle to say about data visualisation? I think before he has a chance to answer, it’s important to offer an apology to the gods (he’ll be there) for flagrantly likening his work, clearly invented for Greek Tragedy, to data visualisation. However, equally humbly I would suggest that it is partly his fault for writing such a universally valid piece about stories — which are such universally applicable means of communication.

In order to figure out what he might have to say about it, we first had to establish the communality between poetry and data visualisation which is roughly covered by the term communication medium as well as by the shared purpose of changing knowledge, belief, opinion or perspective. We also needed to allow ourselves to stray from the path delineated by Aristotle and see it rather as a flexible blueprint and an impulse to focus our thinking on what matters.

However, acknowledging deviations to the original as well as inviting the common ground, we can now try to formulate 7 principles of visual poetics with the aim to support our visual and narrative design process on a conceptual level:

1 Stay close to the theme

The essence of a visualisation is to perceive the story that the information has to tell. Each story is embraced by context. Each context is summarised by theme. Hence, theme is the well and the wall of any communication text. It holds our visualisation together and as such deserves to be the starting point of all consideration and re-visited often. Naturally it will inform the context and our story or stories. Further, it should inform our design to manifest in potential graphical connotations.

2 Script a visual plot

The plot is of central importance to Aristotle, holding the individual pieces of the play not only together, but also giving it a canonical shape with a beginning and an end. Plot in data visualisation is rightfully contested and can by no means be a rule in the original sense. It can only sometimes follow the strict linearity on a general Aristotelian level but can often help the production process by being aware of micro-plots and the user's urge to plot the information field together. As such we could openly rename plot as the tangible fabric the producer passes to the reader to create meaning.

  1. Individual patterns can be read as plots and should be presented linearly and coherently.
  2. Establish a coherent visual plot striving for unity of content, unity of design and where applicable unity of time in order to focus the read.
  3. Aim for a suitable magnitude of expression. Often one of the most difficult points in practice, it is important to find the right spot between complexity and simplicity while neither overstraining nor boring your readers.

3 Design close to your users’ mental models

Characters in Aristotle’s Poetics have the task of engaging the audience through empathy. A data visualisation needs to elicit engagement but can’t draw from the same levels of empathy between a human actor and a human audience. Engagement is achieved by good design.

  1. Appeal to pre-attentive perception through channel effectiveness with visual design.
  2. Tap into existing mental models by using organic or — more broadly — allegorical design.
  3. Be aware of and play to characters or character-like narrative elements with topical design.

4 Constrain your visual diction

Aristotle demands poetic diction — the choice of words — to be clear and focused. The essence of the play needs to be represented unswervingly. Equally, visual diction in data visualisation — the choice of elements, marks and channels — needs to represent the essence of the data and the analysis. Work minimally alongside the data; keep the data-ink ratio high; use colour economically.

5 Set the visual tone in the back

Aristotle wanted the background choir to set the tone of the story — to actively accompany the proceedings on stage. In visualisations the background often takes up a large part of the screen real estate, but it should not often be a large part of the message. Be certain to keep the background in the background while using it to set the tonality of the visual.

6 Use spectacle only to round the corners

Spectacle describes the use of effects in order to heighten excitement and surprise. Blunt in nature and external to the essence, they should be used sparingly to not distract from the core experience. Data visualisation can apply additional stylistic features — redundant animations, notes, transitions — as long as they understand their place in the background; as long as they work towards a rounded experience and not compete for the user’s memory.

7 Refine your user's perspective

Catharsis describes the process of purging the audience’s emotions, aiming for a re-construction to gain a higher level of understanding. The poetic principles can be a road-map to catharsis. With less emotion, but more cognition-focus, a visual data display aims at a similar result that we named refinement. The principles we shamelessly (yet hopefully with respect) re-configured might describe a framework to produce sustainable perspective refinements.

So here we are. 7 principles of visual storytelling, wrapped up in a handy tool-belt — something we have in the under-stairs cupboard to dig out when needed and use when appropriate. No stern rule-book of laws to follow religiously, but rather a box of chocolates to enjoy when we feel like it. Why? Because we're all stories. sort of.


Sources

Some of the fabulous sources (mis?)informing above considerations...

Books · Aristotle, Media

Books · Visualisation

Papers

Posts

Podcasts