Florian Windhager gave the first official lecture of the 3DH series on the subject of Visualization Paradigms. The lecture aimed to give us an overview of the field of information visualization and ran through a number of helpful distinctions.
He started with the simple story of visualization. We have the world, we do observations, these are ordered and then represented by science. We can represent in different ways, with language (text) or with symbolic representations (math) or with pictorial/iconic representations. Some of the common pictorial representations include maps and diagrams.
The phrase “a picture is worth a thousand words” suggests that there is an equivalence between different representations and that one can switch back and forth. Further it suggests that pictures are worth more. A more nuanced view is found in Why a diagram is (sometimes) worth ten thousand words (1987) by Larkin and Simon.
The argument is that is usually presented is that in some cases we have large tables of observations that are difficult for humans to understand. In such cases a visualization can show the data in a way that allows humans, with their natural pattern matching abilities, to quickly grasp the observations and draw appropriate inferences from them. It would be interesting to ask if this is really true and if it can be tested. Why do tables get such a bad name?
A third argument is that images are more similar to certain objects of study and that it therefore takes less cognitive load to process a visualization than other forms of representation.
There are two types of visualization:
- Scientific visualization which is of the objects of science, in particular those objects that are in space. In these cases the visualization can, presumably, resemble to the objects/space of study. Scientific visualizations are naturalistic or isomorphic or mimetic.
- Information visualization which is of something where the space has to be chosen. These are typically of abstractions or of things outside of space like a visualization of some economic data. The designer has to assign dimensions, arrangements, and how to represent objects of study.
In both cases the designer is choosing what they want (us) to see and what will be hidden.
He talked about how visualizations can be used by distinguishing:
- Enabling interactive model based learning and reasoning.
- Providing an overview of details with multiple perspectives.
(To be honest, I think
Is there a visualization of visualizations? He talked about how there are different ways to classify types of visualizations. He recommended the Graphic Continuum which draws on the Data Visualization Catalogue. Moretti in the title of his book mentions three basic types (graphs, maps, and trees).
He then showed us a number of examples that we should look at like The Atlas of Economic Complexity, the Dariah Geo-Browser. He also recommended a good introduction, namely A Tour through the Visualization Zoo by Jeffrey Heer, Michael Bostock, and Vadim Ogievetsky of Stanford University.