The animated transitions paper did bring up an interesting point in how animation can be used in visualization. Although it makes sense as a manner to retain visual consistency between data graphics, I hadn’t thought about using those techniques in terms of creating a sense of causality and attracting attention to the image.

Animated transitions uses animation to aid matching data elements between two visualizations. Animated transitions makes this matching effortless by using motion machinery in the visual cortex.

The paper “Principles of Traditional Animation … ” is a nice summary of the chapters in the book. The 11 design principles of animation looks very convincing but I am sure that they are hard to implement and large number of iteration between the animators and users may be required to have sufficient “entertainment content”.

From the paper “Animated Transitions ….. “, I did get a sense of authors intention, but from the video from their site created doubts about their usefulness and efficacy, Although the paper is nicely written, I still have doubts for the final product.

Chapter 1 presents the principles of animation which could be implemented for motion as encodings or a function affecting encodings. Example of bouncing ball is interesting.

In animated transitions, anticipation, stretch and squash are used in animated transition. Staged transition uses staging additionally. In each subtransition, the use of staggered motion could minimize occlusion. These animated transitions show better accuracy of tracking and perception of changing values as well as preference than direct transition.

Some of the principles seem more concerned with cartoons rather than animation in general. For instance staging – the idea that you can convey one, and only one idea at a time, and that it should be given complete focus. Granted, keeping “clutter” (for some definition of clutter) to a minimum is a good idea, but this seems like an extreme position to take. For instance, if I am doing comparisons in an animation, I want there to be multiple things going on so I can flip back and forth while I compare them.

Squash and stretch can be used for non-rigid objects, as they describe in the text, but (my opinion) if it’s overdone you get something very cartoony.

Exaggeration is synonymous with cartoons, and while it may be worthwhile to consider whether effects are noticeable and effective, there can be such a thing as overdoing them. In the context of animation for the purpose of entertainment, there have been some very good ones without much exaggeration, like Cowboy Bebop, or the Miyazaki films. Incidentally, the ones I’m mentioning here do occasionally have some exaggeration, but usually for brief humorous effect before returning to the main story arc.

Appeal – That is something that is always called for, but almost by definition has no short definition. Oh well.

Arcing motion, followthrough and slow-in / slow-out are actually great ideas for visualization. In cartoons they make things more lifelike and less robotic. For abstract visualizations I think they could be used to at least make the motions more pleasant to look at, and also easier to follow.

Timing can actually be used as a visual channel (speed. jerkiness, etc.) so it’s worth paying attention to as well.

I am interested in statistical data analysis, so my second reading is “Animated Transitions in Statistical Data Graphics”. The video is quite helpful in understanding the content of the paper. My first impression while watching the video is that the animation makes the explanation of statistical graphs, like bar chart, pie chart, easy to understand, especially it helps read focusing on the important part. The representation of the value is simple and clear, but it does not explain the confidence interval in the graphs. I guess the reason is either it requires further knowledge to understand the concept or it is not important. Another useful viewpoint is the change of data schema. Choosing a good view can make explanation different.

Many of the principles laid out by the Disney animators were nicely embodied in the chart transition experiment, some more obvious and some less so. The most obvious one was anticipation, which is providing cues to the viewer that a particular data variable will take an action, in this case transforming its form or moving to a different spot. It provides a good understanding where the variables fit in the various graph formats. also the staging principle is interwoven so that the sequence of graph formats are directed to make the narrative that the data tells us clear. Also form factor principles such as “Squash and Stretch” and “Slow In and Out” are implemented to provide a more fluid and engaging transformation between bars, dots and pies. And not surprisingly, there is a significant level of “appeal” in the data variables – we have seen them performing enjoyable and familiar motions such as jumping, running and switching places. At the end of the presentation, they feel like nerdy incarnations of Mario and Luigi. Overall, I think the way they used animation principles can be of great use to communicate a visual narrative.

I think there is a lot to animation in visualization. The video showing the transition from one representation to another is probably just scratching the surface. The three principles that I saw in that video and with regards to visualization where anticipation, staging, and appeal, although they all, in some way, have a bearing on the final product.

Anticipation is making sure the audience knows what’s going to happen by exaggerating movement. This wasn’t as prevalent in the video, but it would be good to add. For example, if some indication as to the location of a new data point could be given before the animation occured, it would help the viewer track the data.

Staging is probably the most important concept and Tufte would probably file this under “removing visual clutter” although it could add visual clutter. Making sure that the context of the data is absolutely understood is the ultimate goal of a visualization. If the visualization is of birth rates, then adding pictures of babies or baby toys will help the reader understand the visualization much better. In some ways, the staging is a visualization in and of itself. It is sort of like the pre-visualization visualization that gives the viewer the first hint as to what is being conveyed.

Appeal, is probably the second most important part of visualization. If the method of presentation is not appealing to the viewer, it will be that much more difficult for him to understand the data. This could include aesthetic things like picking appropriate color combinations and making sure that there is no aliasing in any lines. It could also be used to describe the method of visualization as in, a bar char is much more appealing than a pie chart.

Those three can be applied to any visualization, when thinking about animation in visualization, all of the other principles should be at play. Things like timing and slow in and out are really important for tracking of data as it moves. I think there is a ton of work that could be done to see how the principles of animation apply to visualization of data.

As a side note, I spent about an hour on youtube watching old Disney shorts and I have to agree with Turetsky, it really is amazing how much of that animation we take for granted and how good it still looks today.