Using sensors to measure control in ping pong
Team: Amanda Rosenberg, Bryan Spence
Type: Sensors-based research class project
When: Spring 2014
Where: IIT Institute of Design
What's the solution?
We developed an initial model of control in a game of ping pong through the combination of data gathered through observation and sensors.
Our goal was to identify control through sensory data. We had a subjective idea of what control meant while watching a game of ping pong, but we wanted to connect what we observed with objective, quantified data.
We found that control could not be quantified by single points, rather by series of points and patterns over the course of multiple games.
How did we get there?
Through a series of experiments, we tracked hit strength and paddle motion with accelerometers, as well as player movement with Microsoft Kinect cameras. We combined these elements with recorded observations from observers and participants while monitoring through video and audio.
To analyze data, we stretched ourselves to use multiple visualization methods like Tableau coordinate graphs, excel formulas, keynote animations, Rhino and accelerometer line graphs. At one point our accelerometer data for one game spanned the length of a hallway.
Takeaways
- Capture more data than you think you’ll need, because you never know what you might discover.
- Synching timestamps and calibrating sensors will make your life easier.
- Experiment with abstract visualizations to see your data through new lenses.
- Make informed decisions on how to slice data.
- Cleaning data will take longer than analyzing it.
We layered the patterns of all games and discovered a general trend in how control is held throughout a game.
