Visualisations & the Process of Abstraction

Visualisations present a reduced view of reality based on abstractions we create.

When information is visualised it can take many forms from diagrams, maps, and information graphics to PPT presentations and wayfinding systems, to name a few. Any of those visualisations can support the processes of learning, teaching, managing information and communication, among others, but the visualisation of information does not always achieve that goal. This is the case of ill-conceived visualisations; these are ambiguous and cluttered representations which obscure knowledge and add confusion, rather than clarity.

When we need to visualise a concrete and tangible problem (e.g. How many legs dogs have?), the visualisation process doesn’t seem to present major obstacles. But when we are dealing with more abstract and unframed problems (e.g. How does the digestive system of dogs work?), the complexity of the visualisation process increases. This post is focuses on the latter type of problems, and describes the process involved in the visualisation of processes, complex situations, and systems.

Process of abstraction

The visualisation of information evolves through a cyclic process of trial and error, resulting from our need to create new ways of understanding and reproducing reality but also improving on key aspects of previous standards. It is a cyclic process in the sense that what today is seen as an effective way to represent information, may not be considered as such in 10 or 20 years time, as societal demands change. What remains constant though is the content (the what) of a visualisation (e.g. how to use a coffee machine, how to navigate a city). In some cases, visualisations have been created with such a balance between what and how, that they are perceived as timeless (e.g. Isotype Language).

1. I know I need to fully understand the problem first (i.e. reason for creating a visualisation), but when do I know I know enough?

Construct Abstract & Functional Understanding. Two types of understanding are required to create effective visualisations (Klein et al., 2007). Abstract understanding describes what is going on in a situation (i.e. our understanding of the problem we need to solve), and functional understanding explains what could be done in that situation (i.e. how our understanding of that problem could be used to find a solution).

2. Then, how do I make the problem manageable?

Create Abstractions & Distortions. Reality presents unstructured information to which we purposely provide a structure in order to construct our own understanding. To create visualisations we simplify reality by creating abstractions of situations, processes, problems or events we want to visually reproduce or communicate. In simple words, visualisations are actually “abstractions from the environment,” or distortions of reality. In other words, to make sense of the world, we first make sense of it by selecting parts, and creating abstractions of them. The process of abstraction involves:

• “Defining continuous processes as discrete steps
• Treating dynamic processes as static
• Treating simultaneous processes as sequential
• Treating complex systems as simple and direct causal mechanisms
• Separating processes that interact
• Treating conditional relationships as universal
• Treating heterogeneous components as homogeneous
• Treating irregular cases as regular ones
• Treating nonlinear functional relationships as linear
• Attending to surface elements rather than deep ones
• Converging on single interpretations rather than multiple interpretations” (Klein et al., 2007)

After we have given reality (e.g. a problem, a situation) an initial structure (NB: not to be confused with the visual structure), we then need to simplify and filter that structure to make it even more manageable and be able to visualise it.

3. How do I organise the content?

Simplify & Filter. As visualisations cannot represent everything, simplification and filtering actions followed the process of abstraction. These actions are undertaken in response to the purpose of the visualisation and the needs of the intended audience. The creator filters relevant from irrelevant content, decides what information should be reproduced and determines the overall structure of the visualisation, e.g. how the information will be organised. Then the creator starts thinking about which would be the adequate visual language and how the selected information can be translated.

4. And what do I do with the organised content?

Select the Right Cues. Creators need to remember that distortion is an intrinsic part of the process to help make sense of situations and translate them into visual language. Distortion does not refer to an alteration of the veracity or authenticity of the content, but it is related to the use of visual metaphors to reproduce reality. Any type of visual representation from a photograph to a drawing is a distortion of reality. This means that each visualisation conveys a distorted view of a situation. The main point is that distortion must be the result of having abstracted the appropriate cues. When we abstract the wrong cues (i.e. we are oversimplifying a situation) both our understanding and resulting visualisations will be also oversimplified. This is the case of visualisations which present a poor or misleading explanation of a problem. To achieve a good visual explanation, the following aspects need to be considered:

• Include key appropriate information
• Display meaningful hierarchical structures of the information
• Respond to the intended audience’s needs
• Apply appropriate visual language

The work by Moles and Janiszewski (1989) elaborates on the process of abstraction providing a scale of iconicity. The scale presents 12 degrees of visual abstraction (from a real object to the mathematical equation representing that object, having in between a realistic photograph of the object, a drawing and a diagrammatic representation of the object) that can be used to reproduce reality.

—-

– Klein, G., Phillips, J. K., Rall, E. L. & Peluso, D. A. (2007) A Data-Frame Theory of Sensemaking. In Expertise Out of Context: Proceedings of the Sixth International Conference on Naturalistic Decision Making (Expertise: Research and Applications Series) (Pensacola Beach, Florida, May 15-17, 2003). Mahwah, NJ: Lawrence Erlbaum Associates, 113-155.
– Moles, A. & Janiszewski, L. (1989) Grafismo funcional. Barcelona: Ed. CEAC.