SenseHub: An integrated web application for sensory analyses

Abstract

This paper discusses the development and evaluation of SenseHub, a web application based on R statistical language, which is specially tailored to answer unprecedented growth in sensory science. Design of experiment, analyses for panel performance, and analyses for twenty-one sensory methods (covering discriminative, descriptive, affective, and optimisation methods) are available in SenseHub. A total of 35 subjects participated in evaluating the performance of SenseHub versus another statistical software on analysing a sensory data. The example dataset was obtained from another study utilising QDA method on 12 perfumes evaluated by 103 panellists using 21 sensory descriptors. In this study, the participants were asked to answer ten questions about data interpretation followed by 11 questions about satisfaction. Time spent for analysis was recorded. Results show that the participants could answer 87.06% questions correctly when using SenseHub while only 76.25% correct answers were achieved when using Minitab (|t|44:94 = 2.36; P < 0.05). Furthermore, the average analysis and interpretation time spent using SenseHub is quicker than Minitab, 11.53 versus 19.73 minutes (|t|53:27 = 3.61; P < 0.05). Finally, SenseHub has higher average satisfaction scores than Minitab in all eleven criteria (two sample t-test, P = 5.70x10-9-5.77 x 10-4). The average satisfaction scores of SenseHub ranged from 84.57 (user interface) to 90.34 (suitability) while the scores of Minitab ranged from 58.20 (understandable) to 69.49 (data import). In short, SenseHub is not only equipped with powerful statistical tools but also considering user experience thus make it promising as an application for sensory analyses.