Search

Conversational LLMs function as black box systems, leaving users guessing about why they see the output they do. This lack of transparency is potentially problematic, especially given concerns around bias and truthfulness. To address this issue, we present an end-to-end prototype-connecting interpretability techniques with user experience design-that seeks to make chatbots more transparent. We begin by showing evidence that a prominent open-source LLM has a "user model": examining the internal state of the system, we can extract data related to a user's age, gender, educational level, and socioeconomic status. Next, we describe the design of a dashboard that accompanies the chatbot interface, displaying this user model in real time. The dashboard can also be used to control the user model and the system's behavior. Finally, we discuss a study in which users conversed with the instrumented system. Our results suggest that users appreciate seeing internal states, which helped them expose biased behavior and increased their sense of control. Participants also made valuable suggestions that point to future directions for both design and machine learning research. The project page and video demo of our TalkTuner system are available at https://bit.ly/talktuner-project-page, Comment: Project page: https://bit.ly/talktuner-project-page, 38 pages, 23 figures

Computational notebooks have become increasingly popular for exploratory data analysis due to their ability to support data exploration and explanation within a single document. Effective documentation for explaining chart findings during the exploration process is essential as it helps recall and share data analysis. However, documenting chart findings remains a challenge due to its time-consuming and tedious nature. While existing automatic methods alleviate some of the burden on users, they often fail to cater to users' specific interests. In response to these limitations, we present InkSight, a mixed-initiative computational notebook plugin that generates finding documentation based on the user's intent. InkSight allows users to express their intent in specific data subsets through sketching atop visualizations intuitively. To facilitate this, we designed two types of sketches, i.e., open-path and closed-path sketch. Upon receiving a user's sketch, InkSight identifies the sketch type and corresponding selected data items. Subsequently, it filters data fact types based on the sketch and selected data items before employing existing automatic data fact recommendation algorithms to infer data facts. Using large language models (GPT-3.5), InkSight converts data facts into effective natural language documentation. Users can conveniently fine-tune the generated documentation within InkSight. A user study with 12 participants demonstrated the usability and effectiveness of InkSight in expressing user intent and facilitating chart finding documentation., Comment: Accepted by VIS23

Transformer models are revolutionizing machine learning, but their inner workings remain mysterious. In this work, we present a new visualization technique designed to help researchers understand the self-attention mechanism in transformers that allows these models to learn rich, contextual relationships between elements of a sequence. The main idea behind our method is to visualize a joint embedding of the query and key vectors used by transformer models to compute attention. Unlike previous attention visualization techniques, our approach enables the analysis of global patterns across multiple input sequences. We create an interactive visualization tool, AttentionViz (demo: http://attentionviz.com), based on these joint query-key embeddings, and use it to study attention mechanisms in both language and vision transformers. We demonstrate the utility of our approach in improving model understanding and offering new insights about query-key interactions through several application scenarios and expert feedback., Comment: 11 pages, 13 figures

Data videos are becoming increasingly popular in society and academia. Yet little is known about how to create endings that strengthen a lasting impression and persuasion. To fulfill the gap, this work aims to develop guidelines for data video endings by drawing inspiration from cinematic arts. To contextualize cinematic endings in data videos, 111 film endings and 105 data video endings are first analyzed to identify four common styles using the framework of ending punctuation marks. We conducted expert interviews (N=11) and formulated 20 guidelines for creating cinematic endings in data videos. To validate our guidelines, we conducted a user study where 24 participants were invited to design endings with and without our guidelines, which are evaluated by experts and the general public. The participants praise the clarity and usability of the guidelines, and results show that the endings with guidelines are perceived to be more understandable, impressive, and reflective.

Visual analytics (VA) systems have been widely used in various application domains. However, VA systems are complex in design, which imposes a serious problem: although the academic community constantly designs and implements new designs, the designs are difficult to query, understand, and refer to by subsequent designers. To mark a major step forward in tackling this problem, we index VA designs in an expressive and accessible way, transforming the designs into a structured format. We first conducted a workshop study with VA designers to learn user requirements for understanding and retrieving professional designs in VA systems. Thereafter, we came up with an index structure VAID to describe advanced and composited visualization designs with comprehensive labels about their analytical tasks and visual designs. The usefulness of VAID was validated through user studies. Our work opens new perspectives for enhancing the accessibility and reusability of professional visualization designs.

Dashboards, which comprise multiple views on a single display, help analyze and communicate multiple perspectives of data simultaneously. However, creating effective and elegant dashboards is challenging since it requires careful and logical arrangement and coordination of multiple visualizations. To solve the problem, we propose a data-driven approach for mining design rules from dashboards and automating dashboard organization. Specifically, we focus on two prominent aspects of the organization: arrangement, which describes the position, size, and layout of each view in the display space; and coordination, which indicates the interaction between pairwise views. We build a new dataset containing 854 dashboards crawled online, and develop feature engineering methods for describing the single views and view-wise relationships in terms of data, encoding, layout, and interactions. Further, we identify design rules among those features and develop a recommender for dashboard design. We demonstrate the usefulness of DMiner through an expert study and a user study. The expert study shows that our extracted design rules are reasonable and conform to the design practice of experts. Moreover, a comparative user study shows that our recommender could help automate dashboard organization and reach human-level performance. In summary, our work offers a promising starting point for design mining visualizations to build recommenders., Comment: in IEEE Transactions on Visualization and Computer Graphics

Analytical dashboards are popular in business intelligence to facilitate insight discovery with multiple charts. However, creating an effective dashboard is highly demanding, which requires users to have adequate data analysis background and be familiar with professional tools, such as Power BI. To create a dashboard, users have to configure charts by selecting data columns and exploring different chart combinations to optimize the communication of insights, which is trial-and-error. Recent research has started to use deep learning methods for dashboard generation to lower the burden of visualization creation. However, such efforts are greatly hindered by the lack of large-scale and high-quality datasets of dashboards. In this work, we propose using deep reinforcement learning to generate analytical dashboards that can use well-established visualization knowledge and the estimation capacity of reinforcement learning. Specifically, we use visualization knowledge to construct a training environment and rewards for agents to explore and imitate human exploration behavior with a well-designed agent network. The usefulness of the deep reinforcement learning model is demonstrated through ablation studies and user studies. In conclusion, our work opens up new opportunities to develop effective ML-based visualization recommenders without beforehand training datasets.

Data visualization is powerful in persuading an audience. However, when it is done poorly or maliciously, a visualization may become misleading or even deceiving. Visualizations give further strength to the dissemination of misinformation on the Internet. The visualization research community has long been aware of visualizations that misinform the audience, mostly associated with the terms "lie" and "deceptive." Still, these discussions have focused only on a handful of cases. To better understand the landscape of misleading visualizations, we open-coded over one thousand real-world visualizations that have been reported as misleading. From these examples, we discovered 74 types of issues and formed a taxonomy of misleading elements in visualizations. We found four directions that the research community can follow to widen the discussion on misleading visualizations: (1) informal fallacies in visualizations, (2) exploiting conventions and data literacy, (3) deceptive tricks in uncommon charts, and (4) understanding the designers' dilemma. This work lays the groundwork for these research directions, especially in understanding, detecting, and preventing them., Comment: To be presented at EuroVis 2022. Appendix available at https://github.com/leoyuholo/bad-vis-browser/blob/master/misinformed_by_visualization_appendix.pdf

Appropriate gestures can enhance message delivery and audience engagement in both daily communication and public presentations. In this paper, we contribute a visual analytic approach that assists professional public speaking coaches in improving their practice of gesture training through analyzing presentation videos. Manually checking and exploring gesture usage in the presentation videos is often tedious and time-consuming. There lacks an efficient method to help users conduct gesture exploration, which is challenging due to the intrinsically temporal evolution of gestures and their complex correlation to speech content. In this paper, we propose GestureLens, a visual analytics system to facilitate gesture-based and content-based exploration of gesture usage in presentation videos. Specifically, the exploration view enables users to obtain a quick overview of the spatial and temporal distributions of gestures. The dynamic hand movements are firstly aggregated through a heatmap in the gesture space for uncovering spatial patterns, and then decomposed into two mutually perpendicular timelines for revealing temporal patterns. The relation view allows users to explicitly explore the correlation between speech content and gestures by enabling linked analysis and intuitive glyph designs. The video view and dynamic view show the context and overall dynamic movement of the selected gestures, respectively. Two usage scenarios and expert interviews with professional presentation coaches demonstrate the effectiveness and usefulness of GestureLens in facilitating gesture exploration and analysis of presentation videos., Comment: 14 pages, 6 figures. A preprint version of a publication at IEEE Transactions on Visualization and Computer Graphics (TVCG), 2022

Data visualizations are created and shared on the web at an unprecedented speed, raising new needs and questions for processing and analyzing visualizations after they have been generated and digitized. However, existing formalisms focus on operating on a single visualization instead of multiple visualizations, making it challenging to perform analysis tasks such as sorting and clustering visualizations. Through a systematic analysis of previous work, we abstract visualization-related tasks into mathematical operators such as union and propose a design space of visualization operations. We realize the design by developing ComputableViz, a library that supports operations on multiple visualization specifications. To demonstrate its usefulness and extensibility, we present multiple usage scenarios concerning processing and analyzing visualization, such as generating visualization embeddings and automatically making visualizations accessible. We conclude by discussing research opportunities and challenges for managing and exploiting the massive visualizations on the web., Comment: 15 pages, 12 figures. In the ACM Conference on Human Factors in Computing Systems (CHI) 2022

With the wide usage of data visualizations, a huge number of Scalable Vector Graphic (SVG)-based visualizations have been created and shared online. Accordingly, there has been an increasing interest in exploring how to retrieve perceptually similar visualizations from a large corpus, since it can benefit various downstream applications such as visualization recommendation. Existing methods mainly focus on the visual appearance of visualizations by regarding them as bitmap images. However, the structural information intrinsically existing in SVG-based visualizations is ignored. Such structural information can delineate the spatial and hierarchical relationship among visual elements, and characterize visualizations thoroughly from a new perspective. This paper presents a structure-aware method to advance the performance of visualization retrieval by collectively considering both the visual and structural information. We extensively evaluated our approach through quantitative comparisons, a user study and case studies. The results demonstrate the effectiveness of our approach and its advantages over existing methods., Comment: Accepted to the 2022 CHI Conference on Human Factors in Computing Systems (CHI 22)

The last decade has witnessed many visual analytics (VA) systems that make successful applications to wide-ranging domains like urban analytics and explainable AI. However, their research rigor and contributions have been extensively challenged within the visualization community. We come in defence of VA systems by contributing two interview studies for gathering critics and responses to those criticisms. First, we interview 24 researchers to collect criticisms the review comments on their VA work. Through an iterative coding and refinement process, the interview feedback is summarized into a list of 36 common criticisms. Second, we interview 17 researchers to validate our list and collect their responses, thereby discussing implications for defending and improving the scientific values and rigor of VA systems. We highlight that the presented knowledge is deep, extensive, but also imperfect, provocative, and controversial, and thus recommend reading with an inclusive and critical eye. We hope our work can provide thoughts and foundations for conducting VA research and spark discussions to promote the research field forward more rigorously and vibrantly., Comment: 9+2 pages, 4 figures. Accepted to IEEE VIS 2022

Data visualizations have been increasingly used in oral presentations to communicate data patterns to the general public. Clear verbal introductions of visualizations to explain how to interpret the visually encoded information are essential to convey the takeaways and avoid misunderstandings. We contribute a series of studies to investigate how to effectively introduce visualizations to the audience with varying degrees of visualization literacy. We begin with understanding how people are introducing visualizations. We crowdsource 110 introductions of visualizations and categorize them based on their content and structures. From these crowdsourced introductions, we identify different introduction strategies and generate a set of introductions for evaluation. We conduct experiments to systematically compare the effectiveness of different introduction strategies across four visualizations with 1,080 participants. We find that introductions explaining visual encodings with concrete examples are the most effective. Our study provides both qualitative and quantitative insights into how to construct effective verbal introductions of visualizations in presentations, inspiring further research in data storytelling., Comment: 12 pages, 5 figures, accepted to IEEE Transaction on Visualization and Graphics

We contribute a deep-learning-based method that assists in designing analytical dashboards for analyzing a data table. Given a data table, data workers usually need to experience a tedious and time-consuming process to select meaningful combinations of data columns for creating charts. This process is further complicated by the need of creating dashboards composed of multiple views that unveil different perspectives of data. Existing automated approaches for recommending multiple-view visualizations mainly build on manually crafted design rules, producing sub-optimal or irrelevant suggestions. To address this gap, we present a deep learning approach for selecting data columns and recommending multiple charts. More importantly, we integrate the deep learning models into a mixed-initiative system. Our model could make recommendations given optional user-input selections of data columns. The model, in turn, learns from provenance data of authoring logs in an offline manner. We compare our deep learning model with existing methods for visualization recommendation and conduct a user study to evaluate the usefulness of the system., Comment: Accepted at the IEEE Visualization Conference (IEEE VIS 2021). 11 pages, 10 figures

Visualizations themselves have become a data format. Akin to other data formats such as text and images, visualizations are increasingly created, stored, shared, and (re-)used with artificial intelligence (AI) techniques. In this survey, we probe the underlying vision of formalizing visualizations as an emerging data format and review the recent advance in applying AI techniques to visualization data (AI4VIS). We define visualization data as the digital representations of visualizations in computers and focus on data visualization (e.g., charts and infographics). We build our survey upon a corpus spanning ten different fields in computer science with an eye toward identifying important common interests. Our resulting taxonomy is organized around WHAT is visualization data and its representation, WHY and HOW to apply AI to visualization data. We highlight a set of common tasks that researchers apply to the visualization data and present a detailed discussion of AI approaches developed to accomplish those tasks. Drawing upon our literature review, we discuss several important research questions surrounding the management and exploitation of visualization data, as well as the role of AI in support of those processes. We make the list of surveyed papers and related material available online at ai4vis.github.io., Comment: Accepted to IEEE TVCG. 20 pages, 12 figures and 8 tables. The associated website is https://ai4vis.github.io

We contribute a method to automate parameter configurations for chart layouts by learning from human preferences. Existing charting tools usually determine the layout parameters using predefined heuristics, producing sub-optimal layouts. People can repeatedly adjust multiple parameters (e.g., chart size, gap) to achieve visually appealing layouts. However, this trial-and-error process is unsystematic and time-consuming, without a guarantee of improvement. To address this issue, we develop Layout Quality Quantifier (LQ2), a machine learning model that learns to score chart layouts from pairwise crowdsourcing data. Combined with optimization techniques, LQ2 recommends layout parameters that improve the charts' layout quality. We apply LQ2 on bar charts and conduct user studies to evaluate its effectiveness by examining the quality of layouts it produces. Results show that LQ2 can generate more visually appealing layouts than both laypeople and baselines. This work demonstrates the feasibility and usages of quantifying human preferences and aesthetics for chart layouts., Comment: Accepted at ACM CHI 2021

GIFs are enjoying increasing popularity on social media as a format for data-driven storytelling with visualization; simple visual messages are embedded in short animations that usually last less than 15 seconds and are played in automatic repetition. In this paper, we ask the question, "What makes a data-GIF understandable?" While other storytelling formats such as data videos, infographics, or data comics are relatively well studied, we have little knowledge about the design factors and principles for "data-GIFs". To close this gap, we provide results from semi-structured interviews and an online study with a total of 118 participants investigating the impact of design decisions on the understandability of data-GIFs. The study and our consequent analysis are informed by a systematic review and structured design space of 108 data-GIFs that we found online. Our results show the impact of design dimensions from our design space such as animation encoding, context preservation, or repetition on viewers' understanding of the GIF's core message. The paper concludes with a list of suggestions for creating more effective Data-GIFs.

We contribute MobileVisFixer, a new method to make visualizations more mobile-friendly. Although mobile devices have become the primary means of accessing information on the web, many existing visualizations are not optimized for small screens and can lead to a frustrating user experience. Currently, practitioners and researchers have to engage in a tedious and time-consuming process to ensure that their designs scale to screens of different sizes, and existing toolkits and libraries provide little support in diagnosing and repairing issues. To address this challenge, MobileVisFixer automates a mobile-friendly visualization re-design process with a novel reinforcement learning framework. To inform the design of MobileVisFixer, we first collected and analyzed SVG-based visualizations on the web, and identified five common mobile-friendly issues. MobileVisFixer addresses four of these issues on single-view Cartesian visualizations with linear or discrete scales by a Markov Decision Process model that is both generalizable across various visualizations and fully explainable. MobileVisFixer deconstructs charts into declarative formats, and uses a greedy heuristic based on Policy Gradient methods to find solutions to this difficult, multi-criteria optimization problem in reasonable time. In addition, MobileVisFixer can be easily extended with the incorporation of optimization algorithms for data visualizations. Quantitative evaluation on two real-world datasets demonstrates the effectiveness and generalizability of our method., Comment: Accepted at IEEE VIS 2020 (Info VIS)

The modulation of voice properties, such as pitch, volume, and speed, is crucial for delivering a successful public speech. However, it is challenging to master different voice modulation skills. Though many guidelines are available, they are often not practical enough to be applied in different public speaking situations, especially for novice speakers. We present VoiceCoach, an interactive evidence-based approach to facilitate the effective training of voice modulation skills. Specifically, we have analyzed the voice modulation skills from 2623 high-quality speeches (i.e., TED Talks) and use them as the benchmark dataset. Given a voice input, VoiceCoach automatically recommends good voice modulation examples from the dataset based on the similarity of both sentence structures and voice modulation skills. Immediate and quantitative visual feedback is provided to guide further improvement. The expert interviews and the user study provide support for the effectiveness and usability of VoiceCoach., Comment: Accepted by CHI '20

Emotions play a key role in human communication and public presentations. Human emotions are usually expressed through multiple modalities. Therefore, exploring multimodal emotions and their coherence is of great value for understanding emotional expressions in presentations and improving presentation skills. However, manually watching and studying presentation videos is often tedious and time-consuming. There is a lack of tool support to help conduct an efficient and in-depth multi-level analysis. Thus, in this paper, we introduce EmoCo, an interactive visual analytics system to facilitate efficient analysis of emotion coherence across facial, text, and audio modalities in presentation videos. Our visualization system features a channel coherence view and a sentence clustering view that together enable users to obtain a quick overview of emotion coherence and its temporal evolution. In addition, a detail view and word view enable detailed exploration and comparison from the sentence level and word level, respectively. We thoroughly evaluate the proposed system and visualization techniques through two usage scenarios based on TED Talk videos and interviews with two domain experts. The results demonstrate the effectiveness of our system in gaining insights into emotion coherence in presentations., Comment: 11 pages, 8 figures. Accepted by IEEE VAST 2019

To enhance the mobility and convenience of the campus community, we designed and implemented the Pulse system, a visual interface for communicating the crowd information to the lay public including campus members and visitors. This is a challenging task which requires analyzing and reconciling the demands and interests for data as well as visual design among diverse target audiences. Through an iterative design progress, we study and address the diverse preferences of the lay audiences, whereby design rationales are distilled. The final prototype combines a set of techniques such as chart junk and redundancy encoding. Initial feedback from a wide audience confirms the benefits and attractiveness of the system.