I'm looking at this presentation at Ted.com, and it uses impressive animated statistics:
Hans Rosling: Stats that reshape your worldview
Does using animated data make understanding statistical data easier for the user? Or are the "old school" bar charts/graphs still better understood/interpreted?
I think animation for abstract images like graphs helps to highlight key parts of the image. Making it visually stand out, and therefore easier to read.
Assuming the following is true.
One could test this by performing tests of readability between two groups of users. One presented with animated graphs and the other static graphs.
Animated graphs are used by a lot of popular websites.
Animation of data showing the change over time is very useful. It shows another dimension of the data in a way that's intuitive: the change in data (over time) is displayed as a change in the graph (over time).
Adding a 3rd dimension to a 2D graph can be done without animation (see 3D graphs) but it becomes much more cluttered and takes more effort or expertise to understand.
A useful feature of these changing graphs is being able to play them as well as being able to manually control them with a slider.
I'm less enthusiastic about animated data that does not animate change over time. I'm not saying it can't be useful, but it without simplicity of time representing time I think it tends to be less intuitive.
The research on the effects of animation is mixed but trends toward the negative.
One of the best summaries of the research on animation as an explanatory technique - not necessarily in statistical graphics though - is by Tversky.
This is a quote from that paper:
"if there are benefits to animation, they should be evident
especially for continuous rather than discrete changes, in particular, for manner of
change and for microsteps, the subtle and intricate timing relations among parts of a
complex system."
Examples of more recent research on animation in bar charts, pie charts, and scatter plots can be found here. Their work builds on Tversky's principles for animation and describes 7 transitions suitable for 'data graphics'.
As with most problems in this area, the suitability of a design solution, e.g., animation, depends on the details of the problem.
Animation:
Here is my personal opinion. Motion design (aka animation) is a skill. The absence of the skill can lead to animations that detract from interpretation of statistical graphics whereas understanding of animation principles and attention to details will improve presentation of data. Take a look at this paper. It describes attempts to improve a well-known animated scientific visualization. The work they describe requires both an understanding of (1) animation principles, (2) the data being animated, and (3) the purpose of the presentation.
