The Dance of Logic and Unpredictability: Examining the Predictability of User Behavior on Visual Analytics Tasks

TL;DR

The paper investigates when user behavior in visual analytics tasks is predictable and how intelligent VA systems can adapt to both routine and non-routine user actions. It surveys case studies on navigation/prefetching and data-interest prediction, demonstrating feasible predictive signals through Markov models and content-based recommendations. An agent-based conceptual framework for human–AI collaboration is proposed, along with discussions of multi-agent systems and explainability to operationalize these partnerships. Ethical and contextual considerations are highlighted, outlining future directions to build more intuitive, efficient, and trustworthy human–machine collaboration in data analysis.

Abstract

The quest to develop intelligent visual analytics (VA) systems capable of collaborating and naturally interacting with humans presents a multifaceted and intriguing challenge. VA systems designed for collaboration must adeptly navigate a complex landscape filled with the subtleties and unpredictabilities that characterize human behavior. However, it is noteworthy that scenarios exist where human behavior manifests predictably. These scenarios typically involve routine actions or present a limited range of choices. This paper delves into the predictability of user behavior in the context of visual analytics tasks. It offers an evidence-based discussion on the circumstances under which predicting user behavior is feasible and those where it proves challenging. We conclude with a forward-looking discussion of the future work necessary to cultivate more synergistic and efficient partnerships between humans and the VA system. This exploration is not just about understanding our current capabilities and limitations in mirroring human behavior but also about envisioning and paving the way for a future where human-machine interaction is more intuitive and productive.

Figures (5)

• Figure 1: The interface and analysis result from Crouser et al. They analyze the analysis behaviors from a series of exercises with 22 trained intelligence analysts crouser2020investigating. Their preliminary analysis suggests that individual differences in locus of control can modulate expert behavior in complex analysis tasks.
• Figure 2: The Tableau interface with a prototype dashboard with an epidemic data set in the fictitious city of Vastopolis, used as the running example in \ref{['sec:va_impact']}. The text displays a map of social media posts with geolocation, a search and filter sidebar, and a bar chart indicating post frequency over three weeks.
• Figure 3: The ForeCache project interface, which visualizes snow levels from NASA MODIS data battle2016dynamic. The authors used observed navigation patterns to predict future interactions and pre-fetch data.
• Figure 4: The interface used by Monadjemi et al. in evaluating their algorithm that observes data exploration, infers the relevance of the other points in the dataset and recommends content to the user monadjemi2022guided.
• Figure 5: The agent-based framework for visual analytics proposed by monadjemi2023human. It adopts terminologies from ai and conceptualizes visual analytics scenarios as interactions (observations and actions) between agents and their environment.