Animation Performance for Time Trends

Since the first appearance of GapMinder , animation, as a way to display trends and patterns when displaying complex data, has gained more popularity than before. It seems that people want to show more and more bizarre animations a la Gapminder, in their presentations, and an increasing number of new products on the market that are trying to copy this visual style.

No doubt, data driven animations in visualization are great. People seem to like it a lot. As a result, powerful animation features have been incorporated into many popular software packages. But does animation allow people to better perceive complex data?

Several researchers from Microsoft Research and Georgia Tech. studied this question in InfoVis 2008 and presented the results in a document entitled “ Animation Effectiveness in Trend Visualization ” (PDF). A short answer to this question - depending on the situation.

Scientists have compared 3 different data visualizations based on historical data from UN countries. They measured their effectiveness in terms of accuracy, speed, and subjective satisfaction with the result.
Animation: A standard storyline with animated bubbles.
Small networks: a static section of the matrix where each point is one and shows the evolution of the bubbles (countries) from the initial to the final position. (bottom right).
Traces: One standard (static) scatter plot diagram on which traces are drawn on top of each other. (bottom left) The


animation leaves no traces, but simply enlivens the transitions.

The study interestingly differentiates between 2 visualization goals: presentation and analytical, and is designed to compare the performance of each of them. Accordingly, the answer varies depending on what the purpose of the visualization was.

Here is a summary of the main findings:

• Small networks have always been more accurate than others.
• Animation was the fastest in the presentation (60-70% faster).
• Animation was the slowest in the analysis (from 50 to 80% slower).
• People found Animation more interesting, and they liked it.

But when asked about what technique they would prefer, there was no clear winner.

So, what did you learn with this? And how can this be translated into the plane of practical recommendations? Here is a summary of the findings of scientists:

• The difference between the goals of presentation and analysis: the same technique may not be equally effective for analysis and presentation. For different purposes, different solutions are needed.
• Animation can be adequate and interesting for presentation purposes: animation is suitable for presentation. In particular, if you want to get the effect of participating in a presentation, animation will help create an empathy effect.
• You should not expect the audience to understand all the details: the study showed that people can get confused very easily, and the overall accuracy of reading data in each case was low, regardless of technique.
• And you will need a good speaker and a minimum of data: a good interaction with the audience greatly depends on the qualities of the speaker. GapMinder works great primarily because its creator is a brilliant speaker. A perfect technique cannot hide the flaws of a bad speaker. In addition, the audience may be slightly overloaded with data. Do not expect the effect to be as good on a lot of data.

The problem of animation goes far beyond just one study. Some researchers predict limited and narrow application, while others are more optimistic.

One area where animation will have advantages is the visualization of transitions between visualization states. Some use animation as a way to convey complex concepts. But the result in this case is controversial. Finally, studying animation in an environment where dynamic data is rendered certainly deserves attention and will require much more research.

Animation is well distributed and perceived on the Internet. Surely designers use animation only when they believe that it will work to achieve a brilliant result. However, the animation has a strong effect on the viewer, and you should try to use this advantage as much as possible. It would be nice to see more research like this in the future.