Understanding how users interact with systems and experience usability issues is vital in Human-Computer Interaction (HCI). Existing approaches often focus on isolated data types, such as interaction logs or subjective reports, providing only a partial view of the user experience. This work introduces AMUSED, a multimodal dataset that integrates user interaction logs, physiological signals, facial emotion features, and self-assessment reports. The dataset includes expert annotations of eleven types of usability smells, creating a rich resource for investigating relationships between usability problems, user behavior, and emotional states. We conducted experiments with 70 participants interacting with three social networks containing usability issues. We then evaluated various machine learning models to assess the feasibility of automatically detecting these issues. The dataset comprises 24 h of user recordings, with over 20,000 user events, such as clicks, scrolls, and input changes. Our analysis reveals that (i) usability smells frequently co-occur, (ii) negative emotions predominate when severe usability issues arise, and (iii) Gradient Boosting models achieve up to 92% accuracy in detecting usability smells, demonstrating the potential for computational methods in automated usability evaluation. Our findings emphasize the value of emotional metrics in HCI research and highlight promising uses of machine learning to automatically detect usability issues.

Usability evaluation is crucial for assessing systems and applications effectiveness and user-friendliness. Traditional methods, such as Cognitive Walkthrough and Usability Smells, often fail to capture emotional dimensions of user experiences. The emergence of facial emotion analysis offers a novel approach to addressing this limitation by providing a more comprehensive understanding of user interactions. This paper introduces “The Emotionality Tool”, an innovative solution that integrates facial emotion analysis into traditional usability evaluation methodologies. The tool features a user-friendly interface for conducting tests, capturing, and analyzing facial emotions, enabling expert evaluators to gain deeper insights into the emotional aspects of user interactions. Case studies demonstrate the tool's effectiveness in uncovering critical usability issues and its potential applications in refining interface designs. “The Emotionality Tool” represents a valuable resource for incorporating emotional metrics into usability evaluation, enriching methodologies and advancing the field of user experience research.

There is a growing concern in the software industries to ensure that systems have usable interfaces, and it is common for the professional designer to be responsible for pursuing this goal. Given the different design processes available in the literature and the different background of professionals, in this paper we seek to understand how the adoption and definition of the design process occurs in the Brazilian software industry. At first, we carried out an online survey, seeking information about professional experiences and the application of design processes. We selected ten survey respondents to be interviewed. We obtained 73 responses with the survey, from which 78% apply a design process, with Design Thinking being the most cited. By conducting the interviews, we found that the processes applied are usually defined according to the project type, while professionals who do not apply processes report a lack of appreciation of the designer’s work by the team. The results highlight differences in how each company proceeds about its performance in the design sector. Overall, our results suggest that design processes bring a substantial positive impact to the projects they are applied to.

Continuous Integration is a critical problem for software maintenance in global projects, compromising companies’ performance, which tends to accumulate a high-resolution time due to the approval process, conflict resolution, tests, and validations. The process of the validation involves the commit description interpretation and can be automated by NLP-mechanisms. This paper presents an intelligent NLP-based approach to evaluate whether the commits can be integrated into a certain software release based only on their descriptions. Our experiments showed an accuracy of 92.9%.