Visualizing temporal sensory data using animated scatterplots, boxplots, and bagplots
The complexity of sensory data can be overwhelming to the uninitiated, particularly when considering the nature of time-related testing (Time Intensity, Temporal Dominance of Sensations, Temporal Order of Sensations, Progressive Profiling, etc.). Researchers attempt to simplify the story told by these rich data sets through graphs, but the tools currently available limit the meaningful and approachable visualizations that can be created. Our project seeks to use the latest web-based means to enable the creation of relevant and informative visualizations which will be easy to create and flexible enough to make sense of complex Temporal Sensory Data.
We visualize data from temporal studies using animated scatterplots. Raw data is overlayed with boxplots and bagplots (Rousseeuw, Ruts & Tukey, 1999) for univariate and bivariate temporal sensory data, respectively, which assists the viewer in understanding the distribution of data, as well as to identify outlying observations. Time can be cycled automatically or manipulated directly to enable both overview and review, facilitating comparison of products. Our application of animated scatterplot and animated boxplot to temporal sensory data is believed to be novel. Our depiction of two-attribute temporal data as bivariate and depicted using animated scatterplot and animated bagplot is also believed to be novel.
Animated visualizations are be demonstrated using bivariate time intensity data in which 13 trained Time Intensity panelists evaluated 4 fruity liqueur products in triplicate on 2 sensory attributes using Single Attribute Time Intensity.
Leonard, A., Plater Findlay, M., Castura, J.C. (2013). Visualizing temporal sensory data using animated scatterplots, boxplots, and bagplots. In 10th Pangborn Sensory Science Symposium. August 11-15. Rio de Janeiro, Brazil.