1AnimationCS 4460/7450 - Information VisualizationApril 2, 2009John StaskoSpring 2006 CS 7450 2Agenda• Animation in InfoVis− How to do it− Where could it be used?2Spring 2006 CS 7450 3Animation• What is it?Spring 2006 CS 7450 4Definition• Animation− Rapid successive display of many display frames where objects change position/appearance gradually so as to create the illusion of continuous movement3Spring 2006 CS 7450 5Application• Where in InfoVis might animation be handy?Spring 2006 CS 7450 6Use• Possibilities− Use time as an option for space, so can show more data (over time)− To draw attention to something− As a visual encoding of particular variable values− To help transition between views4Spring 2006 CS 7450 7Where?• Have we seen animation used in some of the systems/papers we’ve studied so far?Spring 2006 CS 7450 8What We Know• Perception− Animation is a very strong visual attention mechanismIt’s difficult to focus on other items when animation is nearby5Studies about Perception• How do people perceive animations or animated objects?Spring 2006 CS 7450 9User Study• Moticons – Icons with motion• How well do people detect and identify them?• Are they distracting?Spring 2006 CS 7450 10Bartram, Ware & CalvertIJHCS‘036Spring 2006 CS 7450 11Problem• Motivation − Empirical investigation of the effects of moticons as notification mechanism in a peripheral environment• Three experiments − Experiment 1: Detection− Experiment 2: Identification− Experiment 3: DistractionSpring 2006 CS 7450 12Experiment 1: Detection (1/3)• Goal − Color vs Motion− Detection error rates and detection time• Signal cues− Color change: Green, Red− Two motion types: High Amplitude, Low Amplitude• Task− Detect any cue changes while performing a given primary task7Spring 2006 CS 7450 13Participant position and screen widthScreen layoutPrimary task areaStimulus windowNear FarExperiment 1: Detection (2/3)Spring 2006 CS 7450 14Detection error rates Mean detection timeExperiment 1: Detection (3/3)8Spring 2006 CS 7450 15Experiment 2: Identification (1/2)• Goal − In the real world, displays are crowded with multiple colors and shapes− Identify which visual element on the screen changed and thus which event the signal represents• Signal cues− Color change− Shape change− Four motion types: Smooth linear, Jumpy linear, Smooth zoom, Jumpy zoomSpring 2006 CS 7450 16Experiment 2: Identification (2/2)Identification error rates Identification time9Spring 2006 CS 7450 17Experiment 3: Distraction (1/2)• Goal− Evaluate the distraction and irritation properties of moticons in desktop environment under different task conditions− Tasks− Browsing and studying on-line text− Playing FreeCell − Playing Tetris• Motion cues− Linear− Zoom− Blink− TravelSpring 2006 CS 7450 18Experiment 3: Distraction (2/2)Distraction rating by type and frequency10Spring 2006 CS 7450 19Conclusions• Motion is a strong peripheral cue• Useful for searching and identifying things• But it can be distractingAnimation for Transitions• Can animation help “soften the blow” when a view changes?• Preserve context, allow the viewer to track where things wentSpring 2006 CS 7450 2011Suite of Transitions• Developed variety of different transitions and applications• Performed experiments to see how these are perceivedSpring 2006 CS 7450 21Heer & RobertsonTVCG‘07Focus• What types of animation did they use?• How did they use animation?Spring 2006 CS 7450 2212Transition Types• View transformation• Substrate transformation• Filtering• Ordering• Timestep• Visualization change• Data schema changeSpring 2006 CS 7450 23Design Principles• Congruence (mental matching)− Maintain valid data graphics during transitions− Use consistent syntactic-semantic mappings− Respect semantic correspondence− Avoid ambiguity• Apprehension (easily perceivable)− Group similar transitions− Minimize occlusion− Use simple transitions− Use staging for complex transitions− Make transitions as long as needed, but no longerSpring 2006 CS 7450 2413Key Component• Staging− Animation proceeds in stages, not all at once− Varies by animation type and viewSpring 2006 CS 7450 25DynaVis• Implemented in C# and Direct3D graphics• Let’s see it!Spring 2006 CS 7450 26Video14Experiments• 1 – Track object across transitions− Animation beats no-animation, staged animation better than no staging• 2 – Estimate changing values− Animation generally better• Subjective− Staged animation preferred over basic animation preferred over staticSpring 2006 CS 7450 27Telling the Story• Can animation help explain the data?• One traditional use:− Temporal data – Use animation to show changes in timeSpring 2006 CS 7450 2815GapMinder• Company started by Hans Rosling• Focus on world data (by country), much about economics and health• Spotfire-like scatterplot display augmented with animation (animated bubble chart)• Tells a very compelling story with visualizationsSpring 2006 CS 7450 29TED TalksSpring 2006 CS 7450 3020062007Let’s watch16GapMinder• Purchased by Google• Trendalyzer softwareSpring 2006 CS 7450 31Why so Compelling?• Did the animation really add value to the visualizations?• Was it Rosling’s speaking that makes it compelling?Spring 2006 CS 7450 3217Empirical Study• Examine whether animated bubble charts are beneficial for analysis and presentation• Run an experiment to evaluate the effects of animationSpring 2006 CS 7450 33Robertson et alTVCG‘08Visualizations StudiedSpring 2006 CS 7450 34AnimationTracesSmall multiples18Experiment Design• 3 (animation types) x 2 (data size: small & large) x 2 (presentation vs. analysis)− Presentation vs analysis – between subjects− Others – within subjects• Animation has 10-second default time, but user could control time sliderSpring 2006 CS 7450 35Experiment Design• Data− UN data about countries• Tasks− 24 tasks, 1-3 requires answers perSelect 3 countries whose rate of energy consumption was faster than their rate of GDP per capita growthSelect 2 countries with significant decreases in energy consumptionWhich continent had the least changes in GDP per capitaSpring 2006 CS 7450 3619Conditions• Analysis – straightforward, interactive• Presentation− 6 participants at a time− Presenter described a trend relevant to task, but different− No