This study builds on past research to present a domain-specific empirical investigation of artists and math & computer scientists on their respective relationships to, perceptions of, and interactions with data visualization. We conducted a three-phase study utilizing mixed-methods to investigate performance on visual and text representations of data between domains. Our findings evidenced how math & computer scientists are proficient utilizing text representations of data while artists benefit more from visual chart representations. Finally, we present perspectives from artists to gain an understanding of their approach to visual and mathematical tasks. Our findings indicate that artists are especially adept at statistical visual tasks and that development of cognitive skills could be fostered by individuals to potentially benefit visualization task performance.
This case study presents an investigation on explainable artificial intelligence (AI) visualization in business applications. Design guidelines for human-AI interaction are broad and touch on a range of user experiences with AI. Oftentimes, guidelines are not specific to enterprise scenarios with late-stage end users with limited AI knowledge and experience. We present a three-phase study on a visual design of a machine learning (ML) algorithm output. We conducted a user study on an existing design with limited visual AI explanation cues, ran a redesign workshop with various design and data experts, and conducted a reassessment with systematically applied AI explanation guidelines in place. We surface how users with various tech proficiency and AI/ML backgrounds interact with designs and how visual explanation cues increase understanding and effective decision making of users with low AI/ML familiarity. This design process corroborated the application and impact of existing guidelines and surfaced specific design implications for AI explainability within enterprise design.
User telemetry data offers valuable insights into user behavior and goals when interacting with digital products and services. However, analyzing vast amounts of event sequence data can be challenging. In this case study, we apply transformer neural networks to telemetry event sequence data, yielding the UserJourney2Vector model, which treats event sequences similarly to natural language to distill complex user journeys into latent vector representations. The model enables understanding of typical and anomalous user paths, prediction of next actions, and user segmentation based on behavior. We used the model to obtain user clusters, predictive journeys, and cluster statistics, then conducted interviews with digital product experts to assess potential applications and impact of the model output. Experts responded positively to the model’s ability to bolster user personas with data-driven insights and noted integrating model outputs with current practices could augment product design and data interpretation.