Interpreting sequential profiling data via animated box-and-whisker plots

Sara King/ May 11, 2014/ Poster/ 0 comments

Sequential profiling extends traditional descriptive sensory analysis by collecting data at several pre-established times during evaluation. Intensities in product properties are quantified at each time point, permitting product changes to be captured. Evaluation is conducted with panellists who are trained to identify and scale sensory attributes, which removes ambiguity.

We use animated box-and-whisker plots to visualize sequential profiling results. Animated box-and-whisker plots are created using Data-Driven Documents. Essentially animated box-and-whisker plots provide transitions between static boxplots constructed for different time points. Like static box-and-whisker plots, animated box-and-whisker plots allow for outlier detection and provide a useful summary of the distribution of the data. Non-parametric tests can be conducted to facilitate comparisons of products, accompanied by a visual presentation of pairwise product differences, with respect to particular attributes. This information provides the viewer with information that can either confirm or disconfirm a pattern that suggests itself. Animation can either be controlled by the viewer actively or run automatically. The visualization sequence can also be paused to permit a closer inspection.

In this study, trained sensory panel evaluate four snack bars in triplicate using sequential profiling. Animated box-and-whisker plots reveal an outlying observation related to one of the panellists for the attribute Sweetness, with an intensity value so far beyond any other value in the dataset that we conclude it must have been submitted in error. Several other outliers were noted. The visualization is robust in that it is not influenced by such outliers. One products stands out as having an intense Caramelized Flavour and Sweetness, another for evolving Grain Flavour and Nutty Flavour. Animated box-and-whisker plots provide a snapshot views at each timepoint that incorporate univariate comparisons across samples, as well as a sense of intensity changes at adjacent time points.

Plater Findlay, M., Leonard, A., & Castura, J. C. (2014). Interpreting sequential profiling data via animated box-and-whisker plots. SenseAsia 2014: 1st Asian Sensory and Consumer Research Symposium. 11-13 May. Singapore.


Share this Post

Leave a Comment

Your email address will not be published. Required fields are marked *