LEVA: Using Large Language Models to Enhance Visual Analytics.

Visual analytics supports data analysis tasks within complex domain problems. However, due to the richness of data types, visual designs, and interaction designs, users need to recall and process a significant amount of information when they visually analyze data. These challenges emphasize the need for more intelligent visual analytics methods. Large language models have demonstrated the ability to interpret various forms of textual data, offering the potential to facilitate intelligent support for visual analytics. We propose LEVA, a framework that uses large language models to enhance users' VA workflows at multiple stages: onboarding, exploration, and summarization. To support onboarding, we use large language models to interpret visualization designs and view relationships based on system specifications. For exploration, we use large language models to recommend insights based on the analysis of system status and data to facilitate mixed-initiative exploration. For summarization, we present a selective reporting strategy to retrace analysis history through a stream visualization and generate insight reports with the help of large language models. We demonstrate how LEVA can be integrated into existing visual analytics systems. Two usage scenarios and a user study suggest that LEVA effectively aids users in conducting visual analytics.

TransforLearn: Interactive Visual Tutorial for the Transformer Model.

The widespread adoption of Transformers in deep learning, serving as the core framework for numerous large-scale language models, has sparked significant interest in understanding their underlying mechanisms. However, beginners face difficulties in comprehending and learning Transformers due to its complex structure and abstract data representation. We present TransforLearn, the first interactive visual tutorial designed for deep learning beginners and non-experts to comprehensively learn about Transformers. TransforLearn supports interactions for architecture-driven exploration and task-driven exploration, providing insight into different levels of model details and their working processes. It accommodates interactive views of each layer's operation and mathematical formula, helping users to understand the data flow of long text sequences. By altering the current decoder-based recursive prediction results and combining the downstream task abstractions, users can deeply explore model processes. Our user study revealed that the interactions of TransforLearn are positively received. We observe that TransforLearn facilitates users' accomplishment of study tasks and a grasp of key concepts in Transformer effectively.

Visual Explanation for Open-domain Question Answering with BERT.

Open-domain question answering (OpenQA) is an essential but challenging task in natural language processing that aims to answer questions in natural language formats on the basis of large-scale unstructured passages. Recent research has taken the performance of benchmark datasets to new heights, especially when these datasets are combined with techniques for machine reading comprehension based on Transformer models. However, as identified through our ongoing collaboration with domain experts and our review of literature, three key challenges limit their further improvement: (i) complex data with multiple long texts, (ii) complex model architecture with multiple modules, and (iii) semantically complex decision process. In this paper, we present VEQA, a visual analytics system that helps experts understand the decision reasons of OpenQA and provides insights into model improvement. The system summarizes the data flow within and between modules in the OpenQA model as the decision process takes place at the summary, instance and candidate levels. Specifically, it guides users through a summary visualization of dataset and module response to explore individual instances with a ranking visualization that incorporates context. Furthermore, VEQA supports fine-grained exploration of the decision flow within a single module through a comparative tree visualization. We demonstrate the effectiveness of VEQA in promoting interpretability and providing insights into model enhancement through a case study and expert evaluation.

Dashboard Design Patterns.

This paper introduces design patterns for dashboards to inform dashboard design processes. Despite a growing number of public examples, case studies, and general guidelines there is surprisingly little design guidance for dashboards. Such guidance is necessary to inspire designs and discuss tradeoffs in, e.g., screenspace, interaction, or information shown. Based on a systematic review of 144 dashboards, we report on eight groups of design patterns that provide common solutions in dashboard design. We discuss combinations of these patterns in "dashboard genres" such as narrative, analytical, or embedded dashboard. We ran a 2-week dashboard design workshop with 23 participants of varying expertise working on their own data and dashboards. We discuss the application of patterns for the dashboard design processes, as well as general design tradeoffs and common challenges. Our work complements previous surveys and aims to support dashboard designers and researchers in co-creation, structured design decisions, as well as future user evaluations about dashboard design guidelines. Detailed pattern descriptions and workshop material can be found online: https://dashboarddesignpatterns.github.io.

Development and Evaluation of Two Approaches of Visual Sensitivity Analysis to Support Epidemiological Modeling.

Computational modeling is a commonly used technology in many scientific disciplines and has played a noticeable role in combating the COVID-19 pandemic. Modeling scientists conduct sensitivity analysis frequently to observe and monitor the behavior of a model during its development and deployment. The traditional algorithmic ranking of sensitivity of different parameters usually does not provide modeling scientists with sufficient information to understand the interactions between different parameters and model outputs, while modeling scientists need to observe a large number of model runs in order to gain actionable information for parameter optimization. To address the above challenge, we developed and compared two visual analytics approaches, namely: algorithm-centric and visualization-assisted, and visualization-centric and algorithm-assisted. We evaluated the two approaches based on a structured analysis of different tasks in visual sensitivity analysis as well as the feedback of domain experts. While the work was carried out in the context of epidemiological modeling, the two approaches developed in this work are directly applicable to a variety of modeling processes featuring time series outputs, and can be extended to work with models with other types of outputs.

Visualization for epidemiological modelling: challenges, solutions, reflections and recommendations.

We report on an ongoing collaboration between epidemiological modellers and visualization researchers by documenting and reflecting upon knowledge constructs-a series of ideas, approaches and methods taken from existing visualization research and practice-deployed and developed to support modelling of the COVID-19 pandemic. Structured independent commentary on these efforts is synthesized through iterative reflection to develop: evidence of the effectiveness and value of visualization in this context; open problems upon which the research communities may focus; guidance for future activity of this type and recommendations to safeguard the achievements and promote, advance, secure and prepare for future collaborations of this kind. In describing and comparing a series of related projects that were undertaken in unprecedented conditions, our hope is that this unique report, and its rich interactive supplementary materials, will guide the scientific community in embracing visualization in its observation, analysis and modelling of data as well as in disseminating findings. Equally we hope to encourage the visualization community to engage with impactful science in addressing its emerging data challenges. If we are successful, this showcase of activity may stimulate mutually beneficial engagement between communities with complementary expertise to address problems of significance in epidemiology and beyond. See https://ramp-vis.github.io/RAMPVIS-PhilTransA-Supplement/. This article is part of the theme issue 'Technical challenges of modelling real-life epidemics and examples of overcoming these'.

Complex model calibration through emulation, a worked example for a stochastic epidemic model.

Uncertainty quantification is a formal paradigm of statistical estimation that aims to account for all uncertainties inherent in the modelling process of real-world complex systems. The methods are directly applicable to stochastic models in epidemiology, however they have thus far not been widely used in this context. In this paper, we provide a tutorial on uncertainty quantification of stochastic epidemic models, aiming to facilitate the use of the uncertainty quantification paradigm for practitioners with other complex stochastic simulators of applied systems. We provide a formal workflow including the important decisions and considerations that need to be taken, and illustrate the methods over a simple stochastic epidemic model of UK SARS-CoV-2 transmission and patient outcome. We also present new approaches to visualisation of outputs from sensitivity analyses and uncertainty quantification more generally in high input and/or output dimensions.

Words of Estimative Correlation: Studying Verbalizations of Scatterplots.

Natural language and visualization are being increasingly deployed together for supporting data analysis in different ways, from multimodal interaction to enriched data summaries and insights. Yet, researchers still lack systematic knowledge on how viewers verbalize their interpretations of visualizations, and how they interpret verbalizations of visualizations in such contexts. We describe two studies aimed at identifying characteristics of data and charts that are relevant in such tasks. The first study asks participants to verbalize what they see in scatterplots that depict various levels of correlations. The second study then asks participants to choose visualizations that match a given verbal description of correlation. We extract key concepts from responses, organize them in a taxonomy and analyze the categorized responses. We observe that participants use a wide range of vocabulary across all scatterplots, but particular concepts are preferred for higher levels of correlation. A comparison between the studies reveals the ambiguity of some of the concepts. We discuss how the results could inform the design of multimodal representations aligned with the data and analytical tasks, and present a research roadmap to deepen the understanding about visualizations and natural language.

Propagating Visual Designs to Numerous Plots and Dashboards.

In the process of developing an infrastructure for providing visualization and visual analytics (VIS) tools to epidemiologists and modeling scientists, we encountered a technical challenge for applying a number of visual designs to numerous datasets rapidly and reliably with limited development resources. In this paper, we present a technical solution to address this challenge. Operationally, we separate the tasks of data management, visual designs, and plots and dashboard deployment in order to streamline the development workflow. Technically, we utilize: an ontology to bring datasets, visual designs, and deployable plots and dashboards under the same management framework; multi-criteria search and ranking algorithms for discovering potential datasets that match a visual design; and a purposely-design user interface for propagating each visual design to appropriate datasets (often in tens and hundreds) and quality-assuring the propagation before the deployment. This technical solution has been used in the development of the RAMPVIS infrastructure for supporting a consortium of epidemiologists and modeling scientists through visualization.

Revisiting the Modifiable Areal Unit Problem in Deep Traffic Prediction with Visual Analytics.

Deep learning methods are being increasingly used for urban traffic prediction where spatiotemporal traffic data is aggregated into sequentially organized matrices that are then fed into convolution-based residual neural networks. However, the widely known modifiable areal unit problem within such aggregation processes can lead to perturbations in the network inputs. This issue can significantly destabilize the feature embeddings and the predictions - rendering deep networks much less useful for the experts. This paper approaches this challenge by leveraging unit visualization techniques that enable the investigation of many-to-many relationships between dynamically varied multi-scalar aggregations of urban traffic data and neural network predictions. Through regular exchanges with a domain expert, we design and develop a visual analytics solution that integrates 1) a Bivariate Map equipped with an advanced bivariate colormap to simultaneously depict input traffic and prediction errors across space, 2) a Moran's I Scatterplot that provides local indicators of spatial association analysis, and 3) a Multi-scale Attribution View that arranges non-linear dot plots in a tree layout to promote model analysis and comparison across scales. We evaluate our approach through a series of case studies involving a real-world dataset of Shenzhen taxi trips, and through interviews with domain experts. We observe that geographical scale variations have important impact on prediction performances, and interactive visual exploration of dynamically varying inputs and outputs benefit experts in the development of deep traffic prediction models.

Supporting Story Synthesis: Bridging the Gap between Visual Analytics and Storytelling.

Visual analytics usually deals with complex data and uses sophisticated algorithmic, visual, and interactive techniques supporting the analysis. Findings and results of the analysis often need to be communicated to an audience that lacks visual analytics expertise. This requires analysis outcomes to be presented in simpler ways than that are typically used in visual analytics systems. However, not only analytical visualizations may be too complex for target audiences but also the information that needs to be presented. Analysis results may consist of multiple components, which may involve multiple heterogeneous facets. Hence, there exists a gap on the path from obtaining analysis findings to communicating them, within which two main challenges lie: information complexity and display complexity. We address this problem by proposing a general framework where data analysis and result presentation are linked by story synthesis, in which the analyst creates and organises story contents. Unlike previous research, where analytic findings are represented by stored display states, we treat findings as data constructs. We focus on selecting, assembling and organizing findings for further presentation rather than on tracking analysis history and enabling dual (i.e., explorative and communicative) use of data displays. In story synthesis, findings are selected, assembled, and arranged in meaningful layouts that take into account the structure of information and inherent properties of its components. We propose a workflow for applying the proposed conceptual framework in designing visual analytics systems and demonstrate the generality of the approach by applying it to two diverse domains, social media and movement analysis.

LDA Ensembles for Interactive Exploration and Categorization of Behaviors.

We define behavior as a set of actions performed by some actor during a period of time. We consider the problem of analyzing a large collection of behaviors by multiple actors, more specifically, identifying typical behaviors and spotting anomalous behaviors. We propose an approach leveraging topic modeling techniques - LDA (Latent Dirichlet Allocation) Ensembles - to represent categories of typical behaviors by topics that are obtained through topic modeling a behavior collection. When such methods are applied to text in natural languages, the quality of the extracted topics are usually judged based on the semantic relatedness of the terms pertinent to the topics. This criterion, however, is not necessarily applicable to topics extracted from non-textual data, such as action sets, since relationships between actions may not be obvious. We have developed a suite of visual and interactive techniques supporting the construction of an appropriate combination of topics based on other criteria, such as distinctiveness and coverage of the behavior set. Two case studies on analyzing operation behaviors in the security management system and visiting behaviors in an amusement park, and the expert evaluation of the first case study demonstrate the effectiveness of our approach.

VASABI: Hierarchical User Profiles for Interactive Visual User Behaviour Analytics.

User behaviour analytics (UBA) systems offer sophisticated models that capture users' behaviour over time with an aim to identify fraudulent activities that do not match their profiles. Motivated by the challenges in the interpretation of UBA models, this paper presents a visual analytics approach to help analysts gain a comprehensive understanding of user behaviour at multiple levels, namely individual and group level. We take a user-centred approach to design a visual analytics framework supporting the analysis of collections of users and the numerous sessions of activities they conduct within digital applications. The framework is centred around the concept of hierarchical user profiles that are built based on features derived from sessions, as well as on user tasks extracted using a topic modelling approach to summarise and stratify user behaviour. We externalise a series of analysis goals and tasks, and evaluate our methods through use cases conducted with experts. We observe that with the aid of interactive visual hierarchical user profiles, analysts are able to conduct exploratory and investigative analysis effectively, and able to understand the characteristics of user behaviour to make informed decisions whilst evaluating suspicious users and activities.

Broadening Intellectual Diversity in Visualization Research Papers.

Promoting a wider range of contribution types can facilitate healthy growth of the visualization community, while increasing the intellectual diversity of visualization research papers. In this paper, we discuss the importance of contribution types and summarize contribution types that can be meaningful in visualization research. We also propose several concrete next steps we can and should take to ensure a successful launch of the contribution types.

Understanding User Behaviour through Action Sequences: From the Usual to the Unusual.

Action sequences, where atomic user actions are represented in a labelled, timestamped form, are becoming a fundamental data asset in the inspection and monitoring of user behaviour in digital systems. Although the analysis of such sequences is highly critical to the investigation of activities in cyber security applications, existing solutions fail to provide a comprehensive understanding due to the complex semantic and temporal characteristics of these data. This paper presents a visual analytics approach that aims to facilitate a user-involved, multi-faceted decision making process during the identification and the investigation of "unusual" action sequences. We first report the results of the task analysis and domain characterisation process. Then we describe the components of our multi-level analysis approach that comprises of constraint-based sequential pattern mining and semantic distance based clustering, and multi-scalar visualisations of users and their sequences. Finally, we demonstrate the applicability of our approach through a case study that involves tasks requiring effective decision-making by a group of domain experts. Although our solution here is tightly informed by a user-centred, domain-focused design process, we present findings and techniques that are transferable to other applications where the analysis of such sequences is of interest.

Small Multiples with Gaps.

Small multiples enable comparison by providing different views of a single data set in a dense and aligned manner. A common frame defines each view, which varies based upon values of a conditioning variable. An increasingly popular use of this technique is to project two-dimensional locations into a gridded space (e.g. grid maps), using the underlying distribution both as the conditioning variable and to determine the grid layout. Using whitespace in this layout has the potential to carry information, especially in a geographic context. Yet, the effects of doing so on the spatial properties of the original units are not understood. We explore the design space offered by such small multiples with gaps. We do so by constructing a comprehensive suite of metrics that capture properties of the layout used to arrange the small multiples for comparison (e.g. compactness and alignment) and the preservation of the original data (e.g. distance, topology and shape). We study these metrics in geographic data sets with varying properties and numbers of gaps. We use simulated annealing to optimize for each metric and measure the effects on the others. To explore these effects systematically, we take a new approach, developing a system to visualize this design space using a set of interactive matrices. We find that adding small amounts of whitespace to small multiple arrays improves some of the characteristics of 2D layouts, such as shape, distance and direction. This comes at the cost of other metrics, such as the retention of topology. Effects vary according to the input maps, with degree of variation in size of input regions found to be a factor. Optima exist for particular metrics in many cases, but at different amounts of whitespace for different maps. We suggest multiple metrics be used in optimized layouts, finding topology to be a primary factor in existing manually-crafted solutions, followed by a trade-off between shape and displacement. But the rich range of possible optimized layouts leads us to challenge single-solution thinking; we suggest to consider alternative optimized layouts for small multiples with gaps. Key to our work is the systematic, quantified and visual approach to exploring design spaces when facing a trade-off between many competing criteria-an approach likely to be of value to the analysis of other design spaces.

Map LineUps: Effects of spatial structure on graphical inference.

Fundamental to the effective use of visualization as an analytic and descriptive tool is the assurance that presenting data visually provides the capability of making inferences from what we see. This paper explores two related approaches to quantifying the confidence we may have in making visual inferences from mapped geospatial data. We adapt Wickham et al.'s 'Visual Line-up' method as a direct analogy with Null Hypothesis Significance Testing (NHST) and propose a new approach for generating more credible spatial null hypotheses. Rather than using as a spatial null hypothesis the unrealistic assumption of complete spatial randomness, we propose spatially autocorrelated simulations as alternative nulls. We conduct a set of crowdsourced experiments (n=361) to determine the just noticeable difference (JND) between pairs of choropleth maps of geographic units controlling for spatial autocorrelation (Moran's I statistic) and geometric configuration (variance in spatial unit area). Results indicate that people's abilities to perceive differences in spatial autocorrelation vary with baseline autocorrelation structure and the geometric configuration of geographic units. These results allow us, for the first time, to construct a visual equivalent of statistical power for geospatial data. Our JND results add to those provided in recent years by Klippel et al. (2011), Harrison et al. (2014) and Kay & Heer (2015) for correlation visualization. Importantly, they provide an empirical basis for an improved construction of visual line-ups for maps and the development of theory to inform geospatial tests of graphical inference.

Designing Progressive and Interactive Analytics Processes for High-Dimensional Data Analysis.

In interactive data analysis processes, the dialogue between the human and the computer is the enabling mechanism that can lead to actionable observations about the phenomena being investigated. It is of paramount importance that this dialogue is not interrupted by slow computational mechanisms that do not consider any known temporal human-computer interaction characteristics that prioritize the perceptual and cognitive capabilities of the users. In cases where the analysis involves an integrated computational method, for instance to reduce the dimensionality of the data or to perform clustering, such non-optimal processes are often likely. To remedy this, progressive computations, where results are iteratively improved, are getting increasing interest in visual analytics. In this paper, we present techniques and design considerations to incorporate progressive methods within interactive analysis processes that involve high-dimensional data. We define methodologies to facilitate processes that adhere to the perceptual characteristics of users and describe how online algorithms can be incorporated within these. A set of design recommendations and according methods to support analysts in accomplishing high-dimensional data analysis tasks are then presented. Our arguments and decisions here are informed by observations gathered over a series of analysis sessions with analysts from finance. We document observations and recommendations from this study and present evidence on how our approach contribute to the efficiency and productivity of interactive visual analysis sessions involving high-dimensional data.

Visualizing Multiple Variables Across Scale and Geography.

Comparing multiple variables to select those that effectively characterize complex entities is important in a wide variety of domains - geodemographics for example. Identifying variables that correlate is a common practice to remove redundancy, but correlation varies across space, with scale and over time, and the frequently used global statistics hide potentially important differentiating local variation. For more comprehensive and robust insights into multivariate relations, these local correlations need to be assessed through various means of defining locality. We explore the geography of this issue, and use novel interactive visualization to identify interdependencies in multivariate data sets to support geographically informed multivariate analysis. We offer terminology for considering scale and locality, visual techniques for establishing the effects of scale on correlation and a theoretical framework through which variation in geographic correlation with scale and locality are addressed explicitly. Prototype software demonstrates how these contributions act together. These techniques enable multiple variables and their geographic characteristics to be considered concurrently as we extend visual parameter space analysis (vPSA) to the spatial domain. We find variable correlations to be sensitive to scale and geography to varying degrees in the context of energy-based geodemographics. This sensitivity depends upon the calculation of locality as well as the geographical and statistical structure of the variable.