Firs, he defined visualization as computer assisted means to enable insight into data. In research, visual analytics have been a hot topic since 2004. Based on the level of integration of visualization and interaction, visual analytics tools can be divided in the following 3 (or is it 4?) categories:
- level 0: no integration,
- level 1a: visualization of results,
- level 1b: making computational analysis interactive,
- level 2: tight integration.
The last level is the one with most potential for research. He continued by presenting the IVA methodology and the IVA loop. Some remarks about the IVA methodology (and tools for interactive visual analytics): it is needed when the user is faced with too much or too complex data; it should support data exploration, data analysis, hypotheses generation and sense making; it should take into account user interests and task at hand; it should support ‘information drill-down’ (i.e. going from overview to details); and it should offer an interactive and iterative visual dialog. The basic IVA loop consists of two steps: visualization (the computer shows the data to the user) and interaction (the user tells the computer what he/she is interested in). It sounds simple, but the execution of these two steps can quickly get complicated and complex – keep in mind that the process must run in real-time to be interactive.
For more on the topic, see Helvig Hauser’s bibliography.