Be Expressive: proposta de método quantitativo para avaliação de UX utilizando expressões faciais

Abstract

User Experience (UX) is the knowledge field that evaluates the impact of a product, system or service during its interaction with users, aiming at the interaction improvement. In both literature and reports provided by the Software Industry, there are many evaluation methods with distinct characteristics, such as quantitative, qualitative, manual and automatic methods. Within this subject, this work is intended to the development of a automatic quantitative UX evaluation method using facial expression. For that, a search for works related to UX methods that use facial expression analysis and UX methods that use representation of human emotions through facial expressions was accomplished, wherein no quantitative UX evaluation method using facial expressions was found. After identifying the availability of developing a new UX evaluation method, a protocol for executing the experiment intended to verify the level of agreement between the results of two other popular UX methods, EmoCards and User Experience Questionnaire (UEQ), and the developed method, Be Expressive (BeeX), was created. The experiment analysis reported a low level of agreement between Be Expressive and EmoCards (𝜅 ≤ .20) and no level of agreement between EmoCards and UEQ (𝜅 ≤ .0) related to their emotional poles. However, the analyses accomplished provide some findings about the relation between users and UX evaluation methods. Moreover, a scale ranging from -100 to 100 was developed in order to classify the output quantitatively. Lastly, 3 interviews with Human-Computer Interaction and User Experience Professionals were executed in order to verify the availability of using the Be Expressive method. Broadly stating, the opinions obtained were favorable to the usage of Be Expressive. Relating to the final considerations, the need for more research on the usage of Be Expressive was verified in order to understand which types of systems can benefit the most from this method and also the need for investigating the impact of different types of UX methods on the users’ evaluation feedback.