Onboarding a user to a visualization dashboard entails explaining its various components, including the chart types used, the data loaded, and the interactions available. Authoring such an onboarding experience is time-consuming and requires significant knowledge and little guidance on how best to complete this task. Depending on their levels of expertise, end users being onboarded to a new dashboard can be either confused and overwhelmed or disinterested and disengaged. We propose interactive dashboard tours (D-Tours) as semi-automated onboarding experiences that preserve the agency of users with various levels of expertise to keep them interested and engaged. Our interactive tours concept draws from open-world game design to give the user freedom in choosing their path through onboarding. We have implemented the concept in a tool called D-TOUR Prototype, which allows authors to craft custom interactive dashboard tours from scratch or using automatic templates. Automatically generated tours can still be customized to use different media (e.g., video, audio, and highlighting) or new narratives to produce an onboarding experience tailored to an individual user. We demonstrate the usefulness of interactive dashboard tours through use cases and expert interviews. Our evaluation shows that authors found the automation in the D-Tour Prototype helpful and time-saving, and users found the created tours engaging and intuitive. This paper and all supplemental materials are available at https://osf.io/6fbjp/.

Explainable Artificial Intelligence (XAI) enables Artificial Intelligence (AI) to explain its decisions. This holds the promise of making AI more understandable to users, improving interaction, and establishing an adequate level of trust. We tested this claim in the high-risk task of AI-assisted mushroom hunting, where people had to decide whether a mushroom was edible or poisonous. In a between-subjects experiment, 328 visitors of an Austrian media art festival played a tablet-based mushroom hunting game while walking through a highly immersive artificial indoor forest. As part of the game, an artificially intelligent app analyzed photos of the mushrooms they found and recommended classifications. One group saw the AI’s decisions only, while a second group additionally received attribution-based and example-based visual explanations of the AI’s recommendation. The results show that participants with visual explanations outperformed participants without explanations in correct edibility assessments and pick-up decisions. This exhibition-based experiment thus replicated the decision-making results of a previous online study. However, unlike in the previous study, the visual explanations did not significantly affect levels of trust or acceptance measures. In a direct comparison, we consequently discuss the findings in terms of generalizability. Besides the scientific contribution, we discuss the direct impact of conducting XAI experiments in immersive art- and game-based environments in exhibition contexts on visitors and local communities by triggering reflection and awareness for psychological issues of human–AI interaction.

Trust calibration is essential in AI-assisted decision-making. If human users understand the rationale on which an AI model has made a prediction, they can decide whether they consider this prediction reasonable. Especially in high-risk tasks such as mushroom hunting (where a wrong decision may be fatal), it is important that users make correct choices to trust or overrule the AI. Various explainable AI (XAI) methods are currently being discussed as potentially useful for facilitating understanding and subsequently calibrating user trust. So far, however, it remains unclear which approaches are most effective. In this article, the effects of XAI methods on human AI-assisted decision-making in the high-risk task of mushroom picking were tested. For that endeavor, the effects of (i) Grad-CAM attributions, (ii) nearest-neighbor examples, and (iii) network-dissection concepts were compared in a between-subjects experiment with \(N=501\) participants representing end-users of the system. In general, nearest-neighbor examples improved decision correctness the most. However, varying effects for different task items became apparent. All explanations seemed to be particularly effective when they revealed reasons to (i) doubt a specific AI classification when the AI was wrong and (ii) trust a specific AI classification when the AI was correct. Our results suggest that well-established methods, such as Grad-CAM attribution maps, might not be as beneficial to end users as expected and that XAI techniques for use in real-world scenarios must be chosen carefully.

In this work we propose Marjorie, a visual analytics approach to address the challenge of analyzing patients' diabetes data during brief regular appointments with their diabetologists. Designed in consultation with diabetologists, Marjorie uses a combination of visual and algorithmic methods to support the exploration of patterns in the data. Patterns of interest include seasonal variations of the glucose profiles, and non-periodic patterns such as fluctuations around mealtimes or periods of hypoglycemia (i.e., glucose levels below the normal range). We introduce a unique representation of glucose data based on modified horizon graphs and hierarchical clustering of adjacent carbohydrate or insulin entries. Semantic zooming allows the exploration of patterns on different levels of temporal detail. We evaluated our solution in a case study, which demonstrated Marjorie's potential to provide valuable insights into therapy parameters and unfavorable eating habits, among others. The study results and informal feedback collected from target users suggest that Marjorie effectively supports patients and diabetologists in the joint exploration of patterns in diabetes data, potentially enabling more informed treatment decisions. A free copy of this paper and all supplemental materials are available at https://osf.io/34t8c/.

ParaDime is a framework for parametric dimensionality reduction (DR). In parametric DR, neural networks are trained to embed high-dimensional data items in a low-dimensional space while minimizing an objective function. ParaDime builds on the idea that the objective functions of several modern DR techniques result from transformed inter-item relationships. It provides a common interface for specifying these relations and transformations and for defining how they are used within the losses that govern the training process. Through this interface, ParaDime unifies parametric versions of DR techniques such as metric MDS, t-SNE, and UMAP. It allows users to fully customize all aspects of the DR process. We show how this ease of customization makes ParaDime suitable for experimenting with interesting techniques such as hybrid classification/embedding models and supervised DR. This way, ParaDime opens up new possibilities for visualizing high-dimensional data.

Spreadsheet-based tools provide a simple yet effective way of calculating values, which makes them the number-one choice for building and formalizing simple models for budget planning and many other applications. A cell in a spreadsheet holds one specific value and gives a discrete, overprecise view of the underlying model. Therefore, spreadsheets are of limited use when investigating the inherent uncertainties of such models and answering what-if questions. Existing extensions typically require a complex modeling process that cannot easily be embedded in a tabular layout. In Fuzzy Spreadsheet, a cell can hold and display a distribution of values. This integrated uncertainty-handling immediately conveys sensitivity and robustness information. The fuzzification of the cells enables calculations not only with precise values but also with distributions, and probabilities. We conservatively added and carefully crafted visuals to maintain the look and feel of a traditional spreadsheet while facilitating what-if analyses. Given a user-specified reference cell, Fuzzy Spreadsheet automatically extracts and visualizes contextually relevant information, such as impact, uncertainty, and degree of neighborhood, for the selected and related cells. To evaluate its usability and the perceived mental effort required, we conducted a user study. The results show that our approach outperforms traditional spreadsheets in terms of answer correctness, response time, and perceived mental effort in almost all tasks tested.

In this work, we propose an interactive visual approach for the exploration and formation of structural relationships in embeddings of high-dimensional data. These structural relationships, such as item sequences, associations of items with groups, and hierarchies between groups of items, are defining properties of many real-world datasets. Nevertheless, most existing methods for the visual exploration of embeddings treat these structures as second-class citizens or do not take them into account at all. In our proposed analysis workflow, users explore enriched scatterplots of the embedding, in which relationships between items and/or groups are visually highlighted. The original high-dimensional data for single items, groups of items, or differences between connected items and groups are accessible through additional summary visualizations. We carefully tailored these summary and difference visualizations to the various data types and semantic contexts. During their exploratory analysis, users can externalize their insights by setting up additional groups and relationships between items and/or groups. We demonstrate the utility and potential impact of our approach by means of two use cases and multiple examples from various domains.

Understanding user behavior patterns and visual analysis strategies is a long-standing challenge. Existing approaches rely largely on time-consuming manual processes such as interviews and the analysis of observational data. While it is technically possible to capture a history of user interactions and application states, it remains difficult to extract and describe analysis strategies based on interaction provenance. In this article, we propose a novel visual approach to the meta-analysis of interaction provenance. We capture single and multiple user sessions as graphs of high-dimensional application states. Our meta-analysis is based on two different types of two-dimensional embeddings of these high-dimensional states: layouts based on (i) topology and (ii) attribute similarity. We applied these visualization approaches to synthetic and real user provenance data captured in two user studies. From our visualizations, we were able to extract patterns for data types and analytical reasoning strategies.

Abstract Dashboards are used ubiquitously to gain and present insights into data by means of interactive visualizations. To bridge the gap between non‐expert dashboard users and potentially complex datasets and/or visualizations, a variety of onboarding strategies are employed, including videos, narration, and interactive tutorials. We propose a process model for dashboard onboarding that formalizes and unifies such diverse onboarding strategies. Our model introduces the onboarding loop alongside the dashboard usage loop. Unpacking the onboarding loop reveals how each onboarding strategy combines selected building blocks of the dashboard with an onboarding narrative. Specific means are applied to this narration sequence for onboarding, which results in onboarding artifacts that are presented to the user via an interface. We concretize these concepts by showing how our process model can be used to describe a selection of real‐world onboarding examples. Finally, we discuss how our model can serve as an actionable blueprint for developing new onboarding systems.

Classifiers are among the most widely used supervised machine learning algorithms. Many classification models exist, and choosing the right one for a given task is difficult. During model selection and debugging, data scientists need to assess classifiers' performances, evaluate their learning behavior over time, and compare different models. Typically, this analysis is based on single-number performance measures such as accuracy. A more detailed evaluation of classifiers is possible by inspecting class errors. The confusion matrix is an established way for visualizing these class errors, but it was not designed with temporal or comparative analysis in mind. More generally, established performance analysis systems do not allow a combined temporal and comparative analysis of class-level information. To address this issue, we propose ConfusionFlow, an interactive, comparative visualization tool that combines the benefits of class confusion matrices with the visualization of performance characteristics over time. ConfusionFlow is model-agnostic and can be used to compare performances for different model types, model architectures, and/or training and test datasets. We demonstrate the usefulness of ConfusionFlow in a case study on instance selection strategies in active learning. We further assess the scalability of ConfusionFlow and present a use case in the context of neural network pruning.

In problem-solving, a path towards a solutions can be viewed as a sequence of decisions. The decisions, made by humans or computers, describe a trajectory through a high-dimensional representation space of the problem. By means of dimensionality reduction, these trajectories can be visualized in lower-dimensional space. Such embedded trajectories have previously been applied to a wide variety of data, but analysis has focused almost exclusively on the self-similarity of single trajectories. In contrast, we describe patterns emerging from drawing many trajectories—for different initial conditions, end states, and solution strategies—in the same embedding space. We argue that general statements about the problem-solving tasks and solving strategies can be made by interpreting these patterns. We explore and characterize such patterns in trajectories resulting from human and machine-made decisions in a variety of application domains: logic puzzles (Rubik’s cube), strategy games (chess), and optimization problems (neural network training). We also discuss the importance of suitably chosen representation spaces and similarity metrics for the embedding.