Data-Driven Decision-Making for Product Design: A Descriptive Feedback and Physiological Metadata Analysis

This chapter presents the results regarding the data analysis made for studying the user’s perception and experience when giving descriptive feedback within the product design process. It also includes metadata analysis regarding current trends and technology used for gathering body parameters such as temperature, heart rate, pH, sweat, and accelerations to quantify and compare user interactions during the development of a new product. The data analysis tools used were Pandas, Matplotlib, Numpy, spaCy, and the Natural Language Toolkit (NLTK) for sentiment analysis. For exploratory analysis, word clouds were created. A total of 536 sentences were analyzed, with an average length of 12 words per sentence. Among these, 52% of the total words were classified as verbs, nouns, adjectives, or adverbs, providing information regarding user experience, while the remaining words served to link sentences and give structure. The sentiment analysis showcased that 54.8% of the sentences provided a positive or negative connotation regarding the user’s experience. In comparison, 45.2% were neutral sentences, providing no information regarding the user’s perception. It is up to the design team to map and classify these comments into engineering requirements based on their experience. The metadata analysis showed the most studied physiological measurements for human activity recognition (HAR) and how the combination of these variables has been studied to quantify user feedback and improve the design process. Human data collection and study of two or more of these variables is helpful for providing designers with an additional tool for descriptive feedback that could improve decision-making.