arXiv subjects

Arvind Satyanarayan

Publications and source records attributed to Arvind Satyanarayan.

At least 37 records · Page 2Linked to original sources

arXiv2023

Deimos: A Grammar of Dynamic Embodied Immersive Visualisation Morphs and Transitions

We present Deimos, a grammar for specifying dynamic embodied immersive visualisation morphs and transitions. A morph is a collection of animated transitions that are dynamically applied to immersive visualisations at runtime and is conceptually modelled as a state machine. It is comprised of state, transition, and signal specifications. States in a morph are used to generate animation keyframes, with transitions connecting two states together. A transition is controlled by signals, which are composable data streams that can be used to enable embodied interaction techniques. Morphs allow immersive representations of data to transform and change shape through user interaction, facilitating the embodied cognition process. We demonstrate the expressivity of Deimos in an example gallery and evaluate its usability in an expert user study of six immersive analytics researchers. Participants found the grammar to be powerful and expressive, and showed interest in drawing upon Deimos' concepts and ideas in their own research.

cs.HC↗

arXiv2022

Increasing Data Equity Through Accessibility

This position statement is a response to the Office of Science and Technology Policy's Request for Information on "Equitable Data Engagement and Accountability." This response considers data equity specifically for people with disabilities. The RFI asks "how Federal agencies can better support collaboration with other levels of government, civil society, and the research community around the production and use of equitable data." We argue that one critically underserved community in the context of data equity is people with disabilities. Today's tools make it extremely difficult for disabled people to (1) interact with data and data visualizations and (2) take jobs that involve working with and visualizing data. Yet access to such data is increasingly critical, and integral, to engaging with government and civil society. We must change the standards and expectations around data practices to include disabled people and support the research necessary to achieve those goals.

cs.HC↗

arXiv2022

Striking a Balance: Reader Takeaways and Preferences when Integrating Text and Charts

While visualizations are an effective way to represent insights about information, they rarely stand alone. When designing a visualization, text is often added to provide additional context and guidance for the reader. However, there is little experimental evidence to guide designers as to what is the right amount of text to show within a chart, what its qualitative properties should be, and where it should be placed. Prior work also shows variation in personal preferences for charts versus textual representations. In this paper, we explore several research questions about the relative value of textual components of visualizations. 302 participants ranked univariate line charts containing varying amounts of text, ranging from no text (except for the axes) to a written paragraph with no visuals. Participants also described what information they could take away from line charts containing text with varying semantic content. We find that heavily annotated charts were not penalized. In fact, participants preferred the charts with the largest number of textual annotations over charts with fewer annotations or text alone. We also find effects of semantic content. For instance, the text that describes statistical or relational components of a chart leads to more takeaways referring to statistics or relational comparisons than text describing elemental or encoded components. Finally, we find different effects for the semantic levels based on the placement of the text on the chart; some kinds of information are best placed in the title, while others should be placed closer to the data. We compile these results into four chart design guidelines and discuss future implications for the combination of text and charts.

cs.HC↗

arXiv2022

Animated Vega-Lite: Unifying Animation with a Grammar of Interactive Graphics

We present Animated Vega-Lite, a set of extensions to Vega-Lite that model animated visualizations as time-varying data queries. In contrast to alternate approaches for specifying animated visualizations, which prize a highly expressive design space, Animated Vega-Lite prioritizes unifying animation with the language's existing abstractions for static and interactive visualizations to enable authors to smoothly move between or combine these modalities. Thus, to compose animation with static visualizations, we represent time as an encoding channel. Time encodings map a data field to animation keyframes, providing a lightweight specification for animations without interaction. To compose animation and interaction, we also represent time as an event stream; Vega-Lite selections, which provide dynamic data queries, are now driven not only by input events but by timer ticks as well. We evaluate the expressiveness of our approach through a gallery of diverse examples that demonstrate coverage over taxonomies of both interaction and animation. We also critically reflect on the conceptual affordances and limitations of our contribution by interviewing five expert developers of existing animation grammars. These reflections highlight the key motivating role of in-the-wild examples, and identify three central tradeoffs: the language design process, the types of animated transitions supported, and how the systems model keyframes.

cs.HC↗

arXiv2022

Rich Screen Reader Experiences for Accessible Data Visualization

Current web accessibility guidelines ask visualization designers to support screen readers via basic non-visual alternatives like textual descriptions and access to raw data tables. But charts do more than summarize data or reproduce tables; they afford interactive data exploration at varying levels of granularity -- from fine-grained datum-by-datum reading to skimming and surfacing high-level trends. In response to the lack of comparable non-visual affordances, we present a set of rich screen reader experiences for accessible data visualization and exploration. Through an iterative co-design process, we identify three key design dimensions for expressive screen reader accessibility: structure, or how chart entities should be organized for a screen reader to traverse; navigation, or the structural, spatial, and targeted operations a user might perform to step through the structure; and, description, or the semantic content, composition, and verbosity of the screen reader's narration. We operationalize these dimensions to prototype screen-reader-accessible visualizations that cover a diverse range of chart types and combinations of our design dimensions. We evaluate a subset of these prototypes in a mixed-methods study with 13 blind and low vision readers. Our findings demonstrate that these designs help users conceptualize data spatially, selectively attend to data of interest at different levels of granularity, and experience control and agency over their data analysis process. An accessible HTML version of this paper is available at: http://vis.csail.mit.edu/pubs/rich-screen-reader-vis-experiences.

cs.HC↗

arXiv2022

Shared Interest: Measuring Human-AI Alignment to Identify Recurring Patterns in Model Behavior

Saliency methods -- techniques to identify the importance of input features on a model's output -- are a common step in understanding neural network behavior. However, interpreting saliency requires tedious manual inspection to identify and aggregate patterns in model behavior, resulting in ad hoc or cherry-picked analysis. To address these concerns, we present Shared Interest: metrics for comparing model reasoning (via saliency) to human reasoning (via ground truth annotations). By providing quantitative descriptors, Shared Interest enables ranking, sorting, and aggregating inputs, thereby facilitating large-scale systematic analysis of model behavior. We use Shared Interest to identify eight recurring patterns in model behavior, such as cases where contextual features or a subset of ground truth features are most important to the model. Working with representative real-world users, we show how Shared Interest can be used to decide if a model is trustworthy, uncover issues missed in manual analyses, and enable interactive probing.

cs.LG↗

arXiv2022

Embedding Comparator: Visualizing Differences in Global Structure and Local Neighborhoods via Small Multiples

Embeddings mapping high-dimensional discrete input to lower-dimensional continuous vector spaces have been widely adopted in machine learning applications as a way to capture domain semantics. Interviewing 13 embedding users across disciplines, we find comparing embeddings is a key task for deployment or downstream analysis but unfolds in a tedious fashion that poorly supports systematic exploration. In response, we present the Embedding Comparator, an interactive system that presents a global comparison of embedding spaces alongside fine-grained inspection of local neighborhoods. It systematically surfaces points of comparison by computing the similarity of the $k$-nearest neighbors of every embedded object between a pair of spaces. Through case studies across multiple modalities, we demonstrate our system rapidly reveals insights, such as semantic changes following fine-tuning, language changes over time, and differences between seemingly similar models. In evaluations with 15 participants, we find our system accelerates comparisons by shifting from laborious manual specification to browsing and manipulating visualizations.

cs.HC↗

arXiv2021

Teaching Humans When To Defer to a Classifier via Exemplars

Expert decision makers are starting to rely on data-driven automated agents to assist them with various tasks. For this collaboration to perform properly, the human decision maker must have a mental model of when and when not to rely on the agent. In this work, we aim to ensure that human decision makers learn a valid mental model of the agent's strengths and weaknesses. To accomplish this goal, we propose an exemplar-based teaching strategy where humans solve the task with the help of the agent and try to formulate a set of guidelines of when and when not to defer. We present a novel parameterization of the human's mental model of the AI that applies a nearest neighbor rule in local regions surrounding the teaching examples. Using this model, we derive a near-optimal strategy for selecting a representative teaching set. We validate the benefits of our teaching strategy on a multi-hop question answering task using crowd workers and find that when workers draw the right lessons from the teaching stage, their task performance improves, we furthermore validate our method on a set of synthetic experiments.

cs.LG↗

arXiv2021

LMdiff: A Visual Diff Tool to Compare Language Models

While different language models are ubiquitous in NLP, it is hard to contrast their outputs and identify which contexts one can handle better than the other. To address this question, we introduce LMdiff, a tool that visually compares probability distributions of two models that differ, e.g., through finetuning, distillation, or simply training with different parameter sizes. LMdiff allows the generation of hypotheses about model behavior by investigating text instances token by token and further assists in choosing these interesting text instances by identifying the most interesting phrases from large corpora. We showcase the applicability of LMdiff for hypothesis generation across multiple case studies. A demo is available at http://lmdiff.net .

cs.CL↗

arXiv2021

Accessible Visualization via Natural Language Descriptions: A Four-Level Model of Semantic Content

Natural language descriptions sometimes accompany visualizations to better communicate and contextualize their insights, and to improve their accessibility for readers with disabilities. However, it is difficult to evaluate the usefulness of these descriptions, and how effectively they improve access to meaningful information, because we have little understanding of the semantic content they convey, and how different readers receive this content. In response, we introduce a conceptual model for the semantic content conveyed by natural language descriptions of visualizations. Developed through a grounded theory analysis of 2,147 sentences, our model spans four levels of semantic content: enumerating visualization construction properties (e.g., marks and encodings); reporting statistical concepts and relations (e.g., extrema and correlations); identifying perceptual and cognitive phenomena (e.g., complex trends and patterns); and elucidating domain-specific insights (e.g., social and political context). To demonstrate how our model can be applied to evaluate the effectiveness of visualization descriptions, we conduct a mixed-methods evaluation with 30 blind and 90 sighted readers, and find that these reader groups differ significantly on which semantic content they rank as most useful. Together, our model and findings suggest that access to meaningful information is strongly reader-specific, and that research in automatic visualization captioning should orient toward descriptions that more richly communicate overall trends and statistics, sensitive to reader preferences. Our work further opens a space of research on natural language as a data interface coequal with visualization.

cs.HC↗

arXiv2021

DIEL: Interactive Visualization Beyond the Here and Now

Interactive visualization design and research have primarily focused on local data and synchronous events. However, for more complex use cases---e.g., remote database access and streaming data sources---developers must grapple with distributed data and asynchronous events. Currently, constructing these use cases is difficult and time-consuming; developers are forced to operationally program low-level details like asynchronous database querying and reactive event handling. This approach is in stark contrast to modern methods for browser-based interactive visualization, which feature high-level declarative specifications. In response, we present DIEL, a declarative framework that supports asynchronous events over distributed data. Like many declarative visualization languages, DIEL developers need only specify what data they want, rather than procedural steps for how to assemble it; uniquely, DIEL models asynchronous events (e.g., user interactions or server responses) as streams of data that are captured in event logs. To specify the state of a user interface at any time, developers author declarative queries over the data and event logs; DIEL compiles and optimizes a corresponding dataflow graph, and synthesizes necessary low-level distributed systems details. We demonstrate DIEL's performance and expressivity through ex-ample interactive visualizations that make diverse use of remote data and coordination of asynchronous events. We further evaluate DIEL's usability using the Cognitive Dimensions of Notations framework, revealing wins such as ease of change, and compromises such as premature commitments.

cs.DB↗

arXiv2021

Intuitively Assessing ML Model Reliability through Example-Based Explanations and Editing Model Inputs

Interpretability methods aim to help users build trust in and understand the capabilities of machine learning models. However, existing approaches often rely on abstract, complex visualizations that poorly map to the task at hand or require non-trivial ML expertise to interpret. Here, we present two visual analytics modules that facilitate an intuitive assessment of model reliability. To help users better characterize and reason about a model's uncertainty, we visualize raw and aggregate information about a given input's nearest neighbors. Using an interactive editor, users can manipulate this input in semantically-meaningful ways, determine the effect on the output, and compare against their prior expectations. We evaluate our interface using an electrocardiogram beat classification case study. Compared to a baseline feature importance interface, we find that 14 physicians are better able to align the model's uncertainty with domain-relevant factors and build intuition about its capabilities and limitations.

cs.HC↗

arXiv2021

Assessing the Impact of Automated Suggestions on Decision Making: Domain Experts Mediate Model Errors but Take Less Initiative

Automated decision support can accelerate tedious tasks as users can focus their attention where it is needed most. However, a key concern is whether users overly trust or cede agency to automation. In this paper, we investigate the effects of introducing automation to annotating clinical texts--a multi-step, error-prone task of identifying clinical concepts (e.g., procedures) in medical notes, and mapping them to labels in a large ontology. We consider two forms of decision aid: recommending which labels to map concepts to, and pre-populating annotation suggestions. Through laboratory studies, we find that 18 clinicians generally build intuition of when to rely on automation and when to exercise their own judgement. However, when presented with fully pre-populated suggestions, these expert users exhibit less agency: accepting improper mentions, and taking less initiative in creating additional annotations. Our findings inform how systems and algorithms should be designed to mitigate the observed issues.

cs.HC↗

arXiv2021

Beyond Expertise and Roles: A Framework to Characterize the Stakeholders of Interpretable Machine Learning and their Needs

To ensure accountability and mitigate harm, it is critical that diverse stakeholders can interrogate black-box automated systems and find information that is understandable, relevant, and useful to them. In this paper, we eschew prior expertise- and role-based categorizations of interpretability stakeholders in favor of a more granular framework that decouples stakeholders' knowledge from their interpretability needs. We characterize stakeholders by their formal, instrumental, and personal knowledge and how it manifests in the contexts of machine learning, the data domain, and the general milieu. We additionally distill a hierarchical typology of stakeholder needs that distinguishes higher-level domain goals from lower-level interpretability tasks. In assessing the descriptive, evaluative, and generative powers of our framework, we find our more nuanced treatment of stakeholders reveals gaps and opportunities in the interpretability literature, adds precision to the design and comparison of user studies, and facilitates a more reflexive approach to conducting this research.

cs.HC↗

arXiv2021

Viral Visualizations: How Coronavirus Skeptics Use Orthodox Data Practices to Promote Unorthodox Science Online

Controversial understandings of the coronavirus pandemic have turned data visualizations into a battleground. Defying public health officials, coronavirus skeptics on US social media spent much of 2020 creating data visualizations showing that the government's pandemic response was excessive and that the crisis was over. This paper investigates how pandemic visualizations circulated on social media, and shows that people who mistrust the scientific establishment often deploy the same rhetorics of data-driven decision-making used by experts, but to advocate for radical policy changes. Using a quantitative analysis of how visualizations spread on Twitter and an ethnographic approach to analyzing conversations about COVID data on Facebook, we document an epistemological gap that leads pro- and anti-mask groups to draw drastically different inferences from similar data. Ultimately, we argue that the deployment of COVID data visualizations reflect a deeper sociopolitical rift regarding the place of science in public life.

cs.HC↗

arXiv2020

Representing Real-Time Multi-User Collaboration in Visualizations

Establishing common ground and maintaining shared awareness amongst participants is a key challenge in collaborative visualization. For real-time collaboration, existing work has primarily focused on synchronizing constituent visualizations - an approach that makes it difficult for users to work independently, or selectively attend to their collaborators' activity. To address this gap, we introduce a design space for representing synchronous multi-user collaboration in visualizations defined by two orthogonal axes: situatedness, or whether collaborators' interactions are overlaid on or shown outside of a user's view, and specificity, or whether collaborators are depicted through abstract, generic representations or through specific means customized for the given visualization. We populate this design space with a variety of examples including generic and custom synchronized cursors, and user legends that collect these cursors together or reproduce collaborators' views as thumbnails. To build common ground, users can interact with these representations by peeking to take a quick look at a collaborator's view, tracking to follow along with a collaborator in real-time, and forking to independently explore the visualization based on a collaborator's work. We present a reference implementation of a wrapper library that converts interactive Vega-Lite charts into collaborative visualizations. We find that our approach affords synchronous collaboration across an expressive range of visual designs and interaction techniques.

cs.HC↗

arXiv2020

Lyra 2: Designing Interactive Visualizations by Demonstration

Recent graphical interfaces offer direct manipulation mechanisms for authoring visualizations, but are largely restricted to static output. To author interactive visualizations, users must instead turn to textual specification; but, such approaches impose a higher technical burden than their graphical counterparts. To bridge this gap, we introduce interaction design by demonstration: a novel method for authoring interaction techniques via direct manipulation. Users perform an interaction (e.g., button clicks, drags, or key presses) directly on the visualization they are editing. The system interprets this performance using a set of heuristics, and produces suggestions of possible interaction designs. Heuristics account for properties of the interaction (e.g., target and event type) as well as the visualization (e.g., mark and scale types, and multiple views). Interaction design suggestions are displayed as thumbnails; users can preview and test these suggestions, iteratively refine them through additional demonstrations, and finally apply and customize them via property inspectors. To evaluate our approach, we instantiate it in Lyra, an existing visualization design environment. We demonstrate its expressive extent with a gallery of diverse examples, and evaluate its usability through a first-use study and via an analysis of its cognitive dimensions. We find that, in Lyra, interaction design by demonstration enables users to rapidly express a wide range of interactive visualizations.

cs.HC↗

arXiv2020

The Effectiveness of Haptic Properties Under Cognitive Load: An Exploratory Study

With the rise of wearables, haptic interfaces are increasingly favored to communicate information in an ambient manner. Despite this expectation, existing guidelines are developed in studies where the participant's focus is entirely on the haptic task. In this work, we systematically study the cognitive load imposed by properties of a haptic signal. Participants wear a haptic device on their forearm, and are asked to perform a 1-back task. Each experimental condition isolates an individual property of the haptic signal (e.g., amplitude, waveform, rhythm) and participants are asked to identify the gradient of the data. We evaluate each condition across 16 participants, measuring participants' response times, error rates, and qualitative and quantitative surveys (e.g., NASA TLX). Our results indicate that gender and language differences may impact preference for some properties, that participants prefer properties that can be rapidly identified, and that amplitude imposes the lowest cognitive load.

cs.HC↗