Understanding user profiling on social media poses significant challenges due to the intertwined complexities of network structures, user interactions, and the multi-dimensional nature of the data. To reduce visual clutter and offer complementary perspectives for engagingly exploring user profiling within these networks, we propose Social Media Island, an interactive 3D metaphoric visualization system. Our system uses 3D mountain metaphors to visualize user profiling, capturing user influence and activity, while tree metaphors visualize the information forwarding process. To support a flexible scope for users to explore in a more intriguing way, we design various interactions such as cutting the mountain to split out a subset with similarity to some extent for further exploration. By using these 3D visualizations with user interactions, Social Media Island facilitates immersion in the data, the fluid exploration of user influence, topic evolution, and the spread of information. The effectiveness of metaphors is evaluated by user studies, and that of the system is evaluated through two case studies.

With the development of high-performance computers, cloud storage, and advanced sensors, people’s ability to gather complex learning data has greatly improved. However, analyzing these data remains a significant challenge. Especially for spatiotemporal learning data such as eye-tracking and mouse movement, understanding and analyzing these data to identify the learning insights behind them is a difficult task. We propose a visualization platform called “MultiScaleAnalyzer”, which employs hierarchical structure to illustrate spatiotemporal learning data in multiple views. From high-level overviews to detailed analyses, “MultiScaleAnalyzer” provides varying resolutions of data tailored to educators’ need. To demonstrate the platform’s effectiveness, we applied “MultiScaleAnalyzer” to a mathematical word problem-solving dataset, showcasing how the visualization platform facilitates the exploration of student problem-solving patterns and strategies.

Misinformation warnings have become the de facto solution for fighting untruthful messages online. Our study brings forth new understandings as to how users cognitively process two types of warning: general and specific contextual warnings; both are adopted by TikTok and Instagram Reels. Actual TikTok and Instagram users were recruited in this study. In general, we confirm that misinformation warnings indeed aid participants in making sound judgments on the video's veracity. However, general contextual warnings are rated the least effective by participants due to (1) the lack of ability to attract and maintain attention, and (2) the inability to provide relevant information to verify specific video content. Secondly, warnings with "False Information" labels are more effective in helping users to make high-quality accuracy judgments than warnings with "Missing Context" or "Partially False Information" labels, indicating that people prefer affirmative misinformation warnings. Finally, our findings contribute to the evolving scholarship on misinformation warning compliance by casting light on nuanced participant perceptions and behaviors that, if not carefully addressed, may hinder the efforts of misinformation mitigation on social media.

Understanding and comparing the evolution of public opinions on a social media event is important. However, such a task requires summarizing rich semantic information and an in-depth comparison of semantics and dynamics at the same time, which is difficult for the analysis. To tackle these challenges, we propose ContextWing, an interactive visual analytics system to support pair-wise comparison for evolving sequential patterns of contexts between two data streams. The computational model of ContextWing generates dynamic topics and sequential patterns, and characterizes public attention and pair-wise correlations. A novel multi-layer bilateral wing metaphor is designed to intuitively visualizes sequential patterns merged by different contexts to reveal the similarities and differences in both temporal and semantic aspects between two streams. Interactive tools support the selection of a central keyword and its contexts to iteratively generate patterns for a focused exploration. The system supports analysis on both static and streaming settings that enables a wider range of application scenarios. We verify the effectiveness and usability of ContextWing from multiple facets, including three case studies, two expert interviews, and a user study.

The multidimensional nature of spatial data poses a challenge for visualization. In this paper, we introduce Phoenixmap, a simple abstract visualization method to address the issue of visualizing multiple spatial distributions at once. The Phoenixmap approach starts by identifying the enclosed outline of the point collection, then assigns different widths to outline segments according to the segments' corresponding inside regions. Thus, one 2D distribution is represented as an outline with varied thicknesses. Phoenixmap is capable of overlaying multiple outlines and comparing them across categories of objects in a 2D space. We chose heatmap as a benchmark spatial visualization method and conducted user studies to compare performances among Phoenixmap, heatmap, and dot distribution map. Based on the analysis and participant feedback, we demonstrate that Phoenixmap 1) allows users to perceive and compare spatial distribution data efficiently; 2) frees up graphics space with a concise form that can provide visualization design possibilities like overlapping; and 3) provides a good quantitative perceptual estimating capability given the proper legends. Finally, we discuss several possible applications of Phoenixmap and present one visualization of multiple species of birds' active regions in a nature preserve.

We proposed a visual analytics tool to analyze social media posts through machine-learning techniques. Latent Dirichlet Allocation and Named Entities Recognition were used to extract semantic information. 12 topics were identified to interpret semantic meanings and reveal spatiotemporal patterns in the dataset. Our visualization consists of topic bubbles, frequency bar chart, stream graph, fisheye list, massage view, map view, word cloud, and social network graph. All the views are linked together and enhanced by efficient interactions. This paper describes the methodology and visualization design in detail.

As autonomous driving gets closer to be widely applied, it is important to guarantee that traffic signs are recognized correctly. Due to change of light conditions and blockage by some other objects, traffic signs can sometimes be partially corrupted. In this paper, we evaluated how machine learning would respond to different types of image corruption by various degrees. Removing a higher percentage of pixels will gradually harm the recognition accuracy, and removal by blocks causes the biggest harm, which is consistent with human observation. Changing to various colors or a single color doesn't seem to cause significant differences. This study, by building a model for traffic sign recognition and evaluating its robustness to various types of image corruptions, provides insights into corruptions of datasets for machine learning in general, and provide potential concern for applying the well-trained model to a new test set with corruptions, which could be a concern for applying autonomous driving in certain areas.

The computer-aided design process is in a complicated non-linear structure involving selections from a pool of configurations with optimized parameters. In order to understand and improve this decision-making process, this paper conducted a user study on students and expert professionals with more than three years of computer-aided design experience. The study revealed the common design problems and challenges faced by CAD designers. The findings also showed that the design approaches students and expert professionals used were different. Additionally, we found that computer-aided designers expect the system to be able to understand vast quantities of multivariate data, control high-quality products for low costs, manage the knowledge personalization and codification within the company, as well as prevent the design mistakes at the design stages. Our findings may lead to the future development of new approaches to improve the computer-aided design process and close up the gap between student and expert professionals in computer-aided design. This paper provides initial support for this future approach.

This paper proposes an optimized convolutional neural network target recognition algorithm for the problem of low recognition rate of synthetic aperture radar (SAR) target training, under the condition of insufficient tag data, translation, rotation and complexity. In order to overcome the shortage of tag data, the convolutional neural network is initialized with a feature set, obtained by principal component analysis (PCA) unsupervised training. In order to improve the training speed while avoiding overfitting, Rectified Linear Unit (ReLU) function is used as the activation function. In order to enhance robustness and reduce the effect of down sampling on feature representation, this work uses a maximum probability sampling method and normalizes the local contrast of feature after convolution layers. The experimental result shows that, compared with traditional convolutional neural network, this approach achieves a higher recognition rate for SAR target and better robustness to various image deformation and complex background.

This paper introduces a novel visual analytics tool to analyze water contamination data. Bar charts and contour maps are used to inspect the data on an aggregate level. In order to identify both spatial and temporal patterns across all areas, we created a series of contour maps and used isolines to aggregate the readings. We also encoded the contours with color such that newer data were on the warmer spectrum (red/orange) and the older data were on the cooler spectrum (blue/green). Through panning, selecting, brushing, and filtering time scale, the user is able to identify patterns of interest in the data sets and highlight problem areas in the Preserve.

Radial visualization is an important technique to depict serial periodic data. Circle clock design is intuitive to encode 24-hour cyclical data. However, the biggest limitation of the design is the accuracy of reading time points on circle. Dodecagon is another way to represent time series data. We empirically evaluated the effectiveness of circle and dodecagon clock design in perceiving specific points in time. A post-testing interview was also conducted to understand participants’ strategies to read the times. Results show that dodecagon is more accurate than circle in terms of reading time points. Dodecagon was voted as a powerful approach to read the time points and circle was regarded as a better beautiful visualization method.

In this paper, we show that the Shapley–Shubik market game model with production naturally generates an equilibration mechanism that can accommodate price stickiness arising from strategic interactions of firms. Unlike New Keynesian models that show similar price stickiness results, the market game model does not require enforcing menu costs or other additional restraints on price adjustment mechanisms in order to generate price stickiness. As such, we suggest that the market game model can provide a good micro-foundation for macroeconomic analysis. We then explicitly show the relationship between a typical firm’s markup of price over marginal cost and its market share.

This paper presents the design rationale of a color-changing and olfactory scarf to affect people's emotional states in a group environment. The goal of the design is to cheer up depressed individuals or calm down those who are overexcited. Our design uses a heart rate sensor and a skin conductance sensor to detect and recognize emotional information. The scarf will change its color and emit an odor to enhance positive emotions or reduce negative ones. We went through a user-centered design process and discussed different forms of design. Given the wearability and comfort characteristics of scarves, we decided to choose scarves as the solution to regulate emotions.

Many elderly people are living apart from their children. We want to develop a system that can reinforce the connection and emotional feelings between aged parents and their children by combing new digital technologies with soft and warm leather material. Through an iterative design process that includes user interview and observation, design exploration, low-fidelity and high-fidelity prototypes, and usability testing, we present a small digital screen attached to an existing leather wallet of the aged parent to receive and display pictures or videos sent from their adult children. By attaching LINK, the old leather wallet serves as an invisible channel, sharing and connecting beautiful family memories without interrupting the private life of either side.

[COMM]gregater is a visual analytic toolset to analyze and identify the temporal communication patterns and dynamic network structure of a social network in the DinoFun Theme Park of the IEEE VAST 2015 Mini Challenge 2. It synergizes the effectiveness of both graphical and statistical analytics to reveal important communication patterns. This paper introduces an effective way to visualize temporal data and draw basic hypothesis pertaining to the challenge questions.

In systems theory, it is well known that the parameter spaces of dynamical systems are stratified into bifurcation regions, with each supporting a different dynamical solution regime. Some can be stable, with different characteristics, such as monotonic stability, periodic damped stability, or multiperiodic damped stability, and some can be unstable, with different characteristics, such as periodic, multiperiodic, or chaotic unstable dynamics. But in general the existence of bifurcation boundaries is normal and should be expected from most dynamical systems, whether linear or nonlinear. Bifurcation boundaries in parameter space are not evidence of model defect. While existence of such bifurcation boundaries is well known in economic theory, econometricians using macroeconometric models rarely take bifurcation into consideration, when producing policy simulations from macroeconometrics models. Such models are routinely simulated only at the point estimates of the models' parameters.

The 2015 VAST mini challenge conjured up an amusement park - DinoFun World, which owns hundreds of hectares and hosts thousands of visitors every day. With access to movement tracking information of visitors, the task of mini challenge 1 was to identify patterns of attendance and park's activities. Understanding visitor groups and monitoring visitors' activities are vitally important for mangers to provide exceptional service and handle unexpected situations. Facing the large and complex tracking dataset, visual analytic system should not only be able to present the overall situation, but also help operators to analyze information and compare between different patterns.

Enabling users to explore the vast volumes of data from different groups is one of product lifecycle management (PLM)'s goals. PLM must solve such problems as isolated "Islands of Data" and "Island of Automation"; the massive data flow of distanced collaborative design, manufacturing, and management; and the incapability of interpreting and synthesizing data from different perspectives.This paper proposes a new approach from a different perspective: information visualization and visual analytics. An interactive information visualization approach was demonstrated in order to help designers gain insights into massive data and make appropriate decisions. Suggested are possible visualization methods for PLM data- structural visualization, temporal visualization, geospatial visualization, 3D model visualization, and multidimensional visualization. This idea is then demonstrated by a case study of developing an Internet-based information visualization system to visualize the Remote Control Helicopter.