We present a new nonlinear dimensionality reduction method, MAPLE, that enhances UMAP by improving manifold modeling. MAPLE employs a self-supervised learning approach to more efficiently encode low-dimensional manifold geometry. Central to this approach are maximum manifold capacity representations (MMCRs), which help untangle complex manifolds by compressing variances among locally similar data points while amplifying variance among dissimilar data points. This design is particularly effective for high-dimensional data with substantial intra-cluster variance and curved manifold structures, such as biological or image data. Our qualitative and quantitative evaluations demonstrate that MAPLE can produce clearer visual cluster separations and finer subcluster resolution than UMAP while maintaining a tractable computational cost.

Despite the widespread use of Uniform Manifold Approximation and Projection (UMAP), the impact of its stochastic optimization process on the results remains underexplored. We observed that it often produces unstable results where the projections of data points are determined mostly by chance rather than reflecting neighboring structures. To address this limitation, we introduce $(r,d)$-stability to UMAP: a framework that analyzes the stochastic positioning of data points in the projection space. To assess how stochastic elements-specifically, initial projection positions and negative sampling-impact UMAP results, we introduce "ghosts", or duplicates of data points representing potential positional variations due to stochasticity. We define a data point's projection as $(r,d)$-stable if its ghosts perturbed within a circle of radius $r$ in the initial projection remain confined within a circle of radius $d$ for their final positions. To efficiently compute the ghost projections, we develop an adaptive dropping scheme that reduces a runtime up to 60% compared to an unoptimized baseline while maintaining approximately 90% of unstable points. We also present a visualization tool that supports the interactive exploration of the $(r,d)$-stability of data points. Finally, we demonstrate the effectiveness of our framework by examining the stability of projections of real-world datasets and present usage guidelines for the effective use of our framework.

Comparing tensors and identifying their (dis)similar structures is fundamental in understanding the underlying phenomena for complex data. Tensor decomposition methods help analysts extract tensors' essential characteristics and aid in visual analytics for tensors. In contrast to dimensionality reduction (DR) methods designed only for analyzing a matrix (i.e., second-order tensor), existing tensor decomposition methods do not support flexible comparative analysis. To address this analysis limitation, we introduce a new tensor decomposition method, named tensor unified linear comparative analysis (TULCA), by extending its DR counterpart, ULCA, for tensor analysis. TULCA integrates discriminant analysis and contrastive learning schemes for tensor decomposition, enabling flexible comparison of tensors. We also introduce an effective method to visualize a core tensor extracted from TULCA into a set of 2D visualizations. We integrate TULCA's functionalities into a visual analytics interface to support analysts in interpreting and refining the TULCA results. We demonstrate the efficacy of TULCA and the visual analytics interface with computational evaluations and two case studies, including an analysis of log data collected from a supercomputer.

Understanding user interactions in digital systems is essential in analyzing user behaviors and improving system usability. However, a collection of interaction sequences is often large and unstructured, making it challenging to uncover interaction patterns. To address this challenge, we introduce a visual analytics approach that integrates hierarchical clustering and process mining techniques to support analysts in exploring unstructured, large interaction sequence data. Our system employs a tailored dynamic time warping-based similarity measure to enable comparison of interaction sequences. Based on the sequence similarities, we provide stepwise, interactive navigation of clustering results with contextual visual cues for refinement and validation. We further apply process mining to characterize derived clusters. Through these hierarchical clustering and process mining steps, analysts can progressively uncover meaningful interaction patterns while utilizing visual guidance and incorporating domain expertise. We demonstrate our system's effectiveness and applicability through two case studies involving system designers, developers, and domain experts.

Synchronous data-driven storytelling with network visualizations presents significant challenges due to the complexity of real-time manipulation of network components. While existing research addresses asynchronous scenarios, there is a lack of effective tools for live presentations. To address this gap, we developed TangibleNet, a projector-based AR prototype that allows presenters to interact with node-link diagrams using double-sided magnets during live presentations. The design process was informed by interviews with professionals experienced in synchronous data storytelling and workshops with 14 HCI/VIS researchers. Insights from the interviews helped identify key design considerations for integrating physical objects as interactive tools in presentation contexts. The workshops contributed to the development of a design space mapping user actions to interaction commands for node-link diagrams. Evaluation with 12 participants confirmed that TangibleNet supports intuitive interactions and enhances presenter autonomy, demonstrating its effectiveness for synchronous network-based data storytelling.

One advancing machine-learning-based analysis approach for multivariate time-series data is representing data as a third-order tensor and then applying dimensionality reduction (DR) methods. In this work, we introduce a visual analytics method that employs multiple interactive DR methods to support both extraction and interpretation of latent patterns of multivariate time-series data. Our method first allows analysts to select an analysis focus from three axes: instance, variable, and time axes. Then, the method applies a multi-step DR method to produce a 2D scatterplot that depicts latent patterns of the selected axis’s elements (e.g., time points). Afterward, the analysts interactively investigate data groups that appeared in the plot with a DR method designed for comparative analysis. The method can be further applied iteratively to perform more precise and detailed analyses. We implement a prototype system and demonstrate the effectiveness of our method by analyzing supercomputer log data.

Multivariate networks are commonly found in real-world data-driven applications. Uncovering and understanding the relations of interest in multivariate networks is not a trivial task. This paper presents a visual analytics workflow for studying multivariate networks to extract associations between different structural and semantic characteristics of the networks (e.g., what are the combinations of attributes largely relating to the density of a social network?). The workflow consists of a neural-network-based learning phase to classify the data based on the chosen input and output attributes, a dimensionality reduction and optimization phase to produce a simplified set of results for examination, and finally an interpreting phase conducted by the user through an interactive visualization interface. A key part of our design is a composite variable construction step that remodels nonlinear features obtained by neural networks into linear features that are intuitive to interpret. We demonstrate the capabilities of this workflow with multiple case studies on networks derived from social media usage and also evaluate the workflow with qualitative feedback from experts.

Although many dimensionality reduction (DR) techniques employ stochastic methods for computational efficiency, such as negative sampling or stochastic gradient descent, their impact on the projection has been underexplored. In this work, we investigate how such stochasticity affects the stability of projections and present a novel DR technique, GhostUMAP, to measure the pointwise instability of projections. Our idea is to introduce clones of data points, "ghosts", into UMAP’s layout optimization process. Ghosts are designed to be completely passive: they do not affect any others but are influenced by attractive and repulsive forces from the original data points. After a single optimization run, GhostUMAP can capture the projection instability of data points by measuring the variance with the projected positions of their ghosts. We also present a successive halving technique to reduce the computation of GhostUMAP. Our results suggest that Ghost-UMAP can reveal unstable data points with a reasonable computational overhead.

Interaction is critical for data analysis and sensemaking. However, designing interactive physicalizations is challenging as it requires cross-disciplinary knowledge in visualization, fabrication, and electronics. Interactive physicalizations are typically produced in an unstructured manner, resulting in unique solutions for a specific dataset, problem, or interaction that cannot be easily extended or adapted to new scenarios or future physicalizations. To mitigate these challenges, we introduce a computational design pipeline to 3D print network physicalizations with integrated sensing capabilities. Networks are ubiquitous, yet their complex geometry also requires significant engineering considerations to provide intuitive, effective interactions for exploration. Using our pipeline, designers can readily produce network physicalizations supporting selection-the most critical atomic operation for interaction-by touch through capacitive sensing and computational inference. Our computational design pipeline introduces a new design paradigm by concurrently considering the form and interactivity of a physicalization into one cohesive fabrication workflow. We evaluate our approach using (i) computational evaluations, (ii) three usage scenarios focusing on general visualization tasks, and (iii) expert interviews. The design paradigm introduced by our pipeline can lower barriers to physicalization research, creation, and adoption.

Dimensionality reduction (DR) plays a vital role in the visual analysis of high-dimensional data. One main aim of DR is to reveal hidden patterns that lie on intrinsic low-dimensional manifolds. However, DR often overlooks important patterns when the manifolds are distorted or masked by certain influential data attributes. This paper presents a feature learning framework, FEALM, designed to generate a set of optimized data projections for nonlinear DR in order to capture important patterns in the hidden manifolds. These projections produce maximally different nearest-neighbor graphs so that resultant DR outcomes are significantly different. To achieve such a capability, we design an optimization algorithm as well as introduce a new graph dissimilarity measure, named neighbor-shape dissimilarity. Additionally, we develop interactive visualizations to assist comparison of obtained DR results and interpretation of each DR result. We demonstrate FEALM’s effectiveness through experiments and case studies using synthetic and real-world datasets.

Adversarial attacks on a convolutional neural network (CNN)—injecting human-imperceptible perturbations into an input image—could fool a high-performance CNN into making incorrect predictions. The success of adversarial attacks raises serious concerns about the robustness of CNNs, and prevents them from being used in safety-critical applications, such as medical diagnosis and autonomous driving. Our work introduces a visual analytics approach to understanding adversarial attacks by answering two questions: (1) Which neurons are more vulnerable to attacks? and (2) Which image features do these vulnerable neurons capture during the prediction? For the first question, we introduce multiple perturbation-based measures to break down the attacking magnitude into individual CNN neurons and rank the neurons by their vulnerability levels. For the second, we identify image features (e.g., cat ears) that highly stimulate a user-selected neuron to augment and validate the neuron’s responsibility. Furthermore, we support an interactive exploration of a large number of neurons by aiding with hierarchical clustering based on the neurons’ roles in the prediction. To this end, a visual analytics system is designed to incorporate visual reasoning for interpreting adversarial attacks. We validate the effectiveness of our system through multiple case studies as well as feedback from domain experts.

Diffusion tensor imaging (DTI) has been used to study the effects of neurodegenerative diseases on neural pathways, which may lead to more reliable and early diagnosis of these diseases as well as a better understanding of how they affect the brain. We introduce a predictive visual analytics system for studying patient groups based on their labeled DTI fiber tract data and corresponding statistics. The system's machine-learning-augmented interface guides the user through an organized and holistic analysis space, including the statistical feature space, the physical space, and the space of patients over different groups. We use a custom machine learning pipeline to help narrow down this large analysis space and then explore it pragmatically through a range of linked visualizations. We conduct several case studies using DTI and T1-weighted images from the research database of Parkinson's Progression Markers Initiative.

Analyzing air pollution data is challenging as there are various analysis focuses from different aspects: feature (what), space (where), and time (when). As in most geospatial analysis problems, besides high-dimensional features, the temporal and spatial dependencies of air pollution induce the complexity of performing analysis. Machine learning methods, such as dimensionality reduction, can extract and summarize important information of the data to lift the burden of understanding such a complicated environment. In this paper, we present a methodology that utilizes multiple machine learning methods to uniformly explore these aspects. With this methodology, we develop a visual analytic system that supports a flexible analysis workflow, allowing domain experts to freely explore different aspects based on their analysis needs. We demonstrate the capability of our system and analysis workflow supporting a variety of analysis tasks with multiple use cases.

Finding the similarities and differences between groups of datasets is a fundamental analysis task. For high-dimensional data, dimensionality reduction (DR) methods are often used to find the characteristics of each group. However, existing DR methods provide limited capability and flexibility for such comparative analysis as each method is designed only for a narrow analysis target, such as identifying factors that most differentiate groups. This paper presents an interactive DR framework where we integrate our new DR method, called ULCA (unified linear comparative analysis), with an interactive visual interface. ULCA unifies two DR schemes, discriminant analysis and contrastive learning, to support various comparative analysis tasks. To provide flexibility for comparative analysis, we develop an optimization algorithm that enables analysts to interactively refine ULCA results. Additionally, the interactive visualization interface facilitates interpretation and refinement of the ULCA results. We evaluate ULCA and the optimization algorithm to show their efficiency as well as present multiple case studies using real-world datasets to demonstrate the usefulness of this framework.

Many real-world applications involve analyzing time-dependent phenomena, which are intrinsically functional, consisting of curves varying over a continuum (e.g., time). When analyzing continuous data, functional data analysis (FDA) provides substantial benefits, such as the ability to study the derivatives and to restrict the ordering of data. However, continuous data inherently has infinite dimensions, and for a long time series, FDA methods often suffer from high computational costs. The analysis problem becomes even more challenging when updating the FDA results for continuously arriving data. In this paper, we present a visual analytics approach for monitoring and reviewing time series data streamed from a hardware system with a focus on identifying outliers by using FDA. To perform FDA while addressing the computational problem, we introduce new incremental and progressive algorithms that promptly generate the magnitude-shape (MS) plot, which conveys both the functional magnitude and shape outlyingness of time series data. In addition, by using an MS plot in conjunction with an FDA version of principal component analysis, we enhance the analyst's ability to investigate the visually-identified outliers. We illustrate the effectiveness of our approach with two use scenarios using real-world datasets. The resulting tool is evaluated by industry experts using real-world streaming datasets.

Analysis of large, high-dimensional, and heterogeneous datasets is challenging as no one technique is suitable for visualizing and clustering such data in order to make sense of the underlying information. For instance, heterogeneous logs detailing machine repair and maintenance in an organization often need to be analyzed to diagnose errors and identify abnormal patterns, formalize root-cause analyses, and plan preventive maintenance. Such real-world datasets are also beset by issues such as inconsistent and/or missing entries. To conduct an effective diagnosis, it is important to extract and understand patterns from the data with support from analytic algorithms (e.g., finding that certain kinds of machine complaints occur more in the summer) while involving the human-in-the-loop. To address these challenges, we adopt existing techniques for dimensionality reduction (DR) and clustering of numerical, categorical, and text data dimensions, and introduce a visual analytics approach that uses multiple coordinated views to connect DR + clustering results across each kind of the data dimension stated. To help analysts label the clusters, each clustering view is supplemented with techniques and visualizations that contrast a cluster of interest with the rest of the dataset. Our approach assists analysts to make sense of machine maintenance logs and their errors. Then the gained insights help them carry out preventive maintenance. We illustrate and evaluate our approach through use cases and expert studies respectively, and discuss generalization of the approach to other heterogeneous data.

Data-driven problem solving in many real-world applications involves analysis of time-dependent multivariate data, for which dimensionality reduction (DR) methods are often used to uncover the intrinsic structure and features of the data. However, DR is usually applied to a subset of data that is either single-time-point multivariate or univariate time-series, resulting in the need to manually examine and correlate the DR results out of different data subsets. When the number of dimensions is large either in terms of the number of time points or attributes, this manual task becomes too tedious and infeasible. In this paper, we present MulTiDR, a new DR framework that enables processing of time-dependent multivariate data as a whole to provide a comprehensive overview of the data. With the framework, we employ DR in two steps. When treating the instances, time points, and attributes of the data as a 3D array, the first DR step reduces the three axes of the array to two, and the second DR step visualizes the data in a lower-dimensional space. In addition, by coupling with a contrastive learning method and interactive visualizations, our framework enhances analysts' ability to interpret DR results. We demonstrate the effectiveness of our framework with four case studies using real-world datasets.

Understanding and tuning the performance of extreme-scale parallel computing systems demands a streaming approach due to the computational cost of applying offline algorithms to vast amounts of performance log data. Analyzing large streaming data is challenging because the rate of receiving data and limited time to comprehend data make it difficult for the analysts to sufficiently examine the data without missing important changes or patterns. To support streaming data analysis, we introduce a visual analytic framework comprising of three modules: data management, analysis, and interactive visualization. The data management module collects various computing and communication performance metrics from the monitored system using streaming data processing techniques and feeds the data to the other two modules. The analysis module automatically identifies important changes and patterns at the required latency. In particular, we introduce a set of online and progressive analysis methods for not only controlling the computational costs but also helping analysts better follow the critical aspects of the analysis results. Finally, the interactive visualization module provides the analysts with a coherent view of the changes and patterns in the continuously captured performance data. Through a multi-faceted case study on performance analysis of parallel discrete-event simulation, we demonstrate the effectiveness of our framework for identifying bottlenecks and locating outliers.

Dimensionality reduction (DR) methods are commonly used for analyzing and visualizing multidimensional data. However, when data is a live streaming feed, conventional DR methods cannot be directly used because of their computational complexity and inability to preserve the projected data positions at previous time points. In addition, the problem becomes even more challenging when the dynamic data records have a varying number of dimensions as often found in real-world applications. This paper presents an incremental DR solution. We enhance an existing incremental PCA method in several ways to ensure its usability for visualizing streaming multidimensional data. First, we use geometric transformation and animation methods to help preserve a viewer's mental map when visualizing the incremental results. Second, to handle data dimension variants, we use an optimization method to estimate the projected data positions, and also convey the resulting uncertainty in the visualization. We demonstrate the effectiveness of our design with two case studies using real-world datasets.

Dimensionality reduction (DR) is frequently used for analyzing and visualizing high-dimensional data as it provides a good first glance of the data. However, to interpret the DR result for gaining useful insights from the data, it would take additional analysis effort such as identifying clusters and understanding their characteristics. While there are many automatic methods (e.g., density-based clustering methods) to identify clusters, effective methods for understanding a cluster's characteristics are still lacking. A cluster can be mostly characterized by its distribution of feature values. Reviewing the original feature values is not a straightforward task when the number of features is large. To address this challenge, we present a visual analytics method that effectively highlights the essential features of a cluster in a DR result. To extract the essential features, we introduce an enhanced usage of contrastive principal component analysis (cPCA). Our method, called ccPCA (contrasting clusters in PCA), can calculate each feature's relative contribution to the contrast between one cluster and other clusters. With ccPCA, we have created an interactive system including a scalable visualization of clusters' feature contributions. We demonstrate the effectiveness of our method and system with case studies using several publicly available datasets.

Machine learning for data-driven diagnosis has been actively studied in medicine to provide better healthcare. Supporting analysis of a patient cohort similar to a patient under treatment is a key task for clinicians to make decisions with high confidence. However, such analysis is not straightforward due to the characteristics of medical records: high dimensionality, irregularity in time, and sparsity. To address this challenge, we introduce a method for similarity calculation of medical records. Our method employs event and sequence embeddings. While we use an autoencoder for the event embedding, we apply its variant with the self-attention mechanism for the sequence embedding. Moreover, in order to better handle the irregularity of data, we enhance the self-attention mechanism with consideration of different time intervals. We have developed a visual analytics system to support comparative studies of patient records. To make a comparison of sequences with different lengths easier, our system incorporates a sequence alignment method. Through its interactive interface, the user can quickly identify patients of interest and conveniently review both the temporal and multivariate aspects of the patient records. We demonstrate the effectiveness of our design and system with case studies using a real-world dataset from the neonatal intensive care unit of UC Davis.

There is a growing trend of applying machine learning methods to medical datasets in order to predict patients’ future status. Although some of these methods achieve high performance, challenges still exist in comparing and evaluating different models through their interpretable information. Such analytics can help clinicians improve evidence-based medical decision making. In this work, we develop a visual analytics system that compares multiple models’ prediction criteria and evaluates their consistency. With our system, users can generate knowledge on different models’ inner criteria and how confidently we can rely on each model’s prediction for a certain patient. Through a case study of a publicly available clinical dataset, we demonstrate the effectiveness of our visual analytics system to assist clinicians and researchers in comparing and quantitatively evaluating different machine learning methods.

Large, complex networks are commonly found in many application domains, such as sociology, biology, and software engineering. Analyzing such networks can be a non-trivial task, as it often takes many interactions to derive a finding. It is thus beneficial to capture and summarize the important steps in an analysis. This provenance would then effectively support recalling, reusing, reproducing, and sharing the analysis process and results. However, the provenance of analyzing a large, complex network would often be a long interaction record. To automatically compose a concise visual summarization of network analysis provenance, we introduce a ranking model together with a reduction algorithm. The model identifies and orders important interactions used in the network analysis. Based on this model, our algorithm is able to minimize the provenance, while still preserving all the essential steps for recalling and sharing the analysis process and results. We create a prototype system demonstrating the effectiveness of our model and algorithm with two usage scenarios.

The overall efficiency of an extreme-scale supercomputer largely relies on the performance of its network interconnects. Several of the state of the art supercomputers use networks based on the increasingly popular Dragonfly topology. It is crucial to study the behavior and performance of different parallel applications running on Dragonfly networks in order to make optimal system configurations and design choices, such as job scheduling and routing strategies. However, in order to study these temporal network behavior, we would need a tool to analyze and correlate numerous sets of multivariate time-series data collected from the Dragonfly’s multi-level hierarchies. This paper presents such a tool–a visual analytics system–that uses the Dragonfly network to investigate the temporal behavior and optimize the communication performance of a supercomputer. We coupled interactive visualization with time-series analysis methods to help reveal hidden patterns in the network behavior with respect to different parallel applications and system configurations. Our system also provides multiple coordinated views for connecting behaviors observed at different levels of the network hierarchies, which effectively helps visual analysis tasks. We demonstrate the effectiveness of the system with a set of case studies. Our system and findings can not only help improve the communication performance of supercomputing applications, but also the network performance of next-generation supercomputers.

Neuroscientists study brain functional connectivity in order to obtain a deeper understanding of how the brain functions. Current studies are mainly based on analyzing the averaged brain connectivity of a group (or groups) due to the high complexity of the collected data in terms of dimensionality, variability, and volume. While it is more desirable for the researchers to explore the potential variability between individual subjects or groups, a data analysis solution meeting this need is absent. In this paper, we present the design and capabilities of such a visual analytics system, which enables neuroscientists to visually compare the differences of brain networks between individual subjects as well as group averages, to explore a large dataset and examine sub-groups of participants that may not have been expected a priori to be of interest, to review detailed information as needed, and to manipulate the data and views to fit their analytical needs with easy interactions. We demonstrate the utility and strengths of this system with case studies using a representative functional connectivity dataset.

As virtual reality (VR) hardware technology becomes more mature and affordable, it is timely to develop visualization applications making use of such technology. How to interact with data in an immersive 3D space is both an interesting and challenging problem, demanding more research investigations. In this paper, we present a gesture input system for graph visualization in a stereoscopic 3D space. We compare desktop mouse input with gesture input with bare hands for performing a set of tasks on graphs. Our study results indicate that users are able to effortlessly manipulate and analyze graphs using gesture input. Furthermore, the results also show that using gestures is more efficient when exploring the complicated graph.