arXiv subjects

Arvind Satyanarayan

Publications and source records attributed to Arvind Satyanarayan.

At least 19 recordsLinked to original sources

arXiv2026

"Here Be Sharks!": Enhancing Scientific Communication and Analysis through Authoring Interactivity

We report on a case study for designing authoring environments for interactive visualizations to enhance scientific work. We conducted a workshop and prototype review with a group of marine biologists. When it came to visualizing their data, participants identified challenges in conveying their research accurately and completely as well as and analyzing it with ease. Based on our findings, authoring environments should consider the context of scientific work aspects such as domain expertise, collaboration culture, and publication traditions. We propose consideration of non-traditional programming languages and environments for scientific work, discuss ways of facilitating interactive visualizations for scientists, and examine ways of meaningfully integrating AI. The critiques, artifacts, and reactions from the scientists along with our analysis and discussion inform how computational tools should be designed for scientific work.

cs.HC↗

arXiv2026

Creo: From One-Shot Image Generation to Progressive, Co-Creative Ideation

Text-to-image (T2I) systems enable rapid generation of high-fidelity imagery but are misaligned with how visual ideas develop. T2I systems generate outputs that make implicit visual decisions on behalf of the user, often introduce fine-grained details that can anchor users prematurely and limit their ability to keep options open early on, and cause unintended changes during editing that are difficult to correct and reduce users' sense of control. To address these concerns, we present Creo, a multi-stage T2I system that scaffolds image generation by progressing from rough sketches to high-resolution outputs, exposing intermediary abstractions where users can make incremental changes. Sketch-like abstractions invite user editing and allow users to keep design options open when ideas are still forming due to their provisional nature. Each stage in Creo can be modified with manual changes and AI-assisted operations, enabling fine-grained, step-wise control through a locking mechanism that preserves prior decisions so subsequent edits affect only specified regions or attributes. Users remain in the loop, making and verifying decisions across stages, while the system applies diffs instead of regenerating full images, reducing drift as fidelity increases. A comparative study with a one-shot baseline shows that participants felt stronger ownership over Creo outputs, as they were able to trace their decisions in building up the image. Furthermore, embedding-based analysis indicates that Creo outputs are less homogeneous than one-shot results. These findings suggest that multi-stage generation, combined with intermediate control and decision locking, is a key design principle for improving controllability, user agency, creativity, and output diversity in generative systems.

cs.HC↗

arXiv2025

Semantic Regexes: Auto-Interpreting LLM Features with a Structured Language

Automated interpretability aims to translate large language model (LLM) features into human understandable descriptions. However, natural language feature descriptions can be vague, inconsistent, and require manual relabeling. In response, we introduce semantic regexes, structured language descriptions of LLM features. By combining primitives that capture linguistic and semantic patterns with modifiers for contextualization, composition, and quantification, semantic regexes produce precise and expressive feature descriptions. Across quantitative benchmarks and qualitative analyses, semantic regexes match the accuracy of natural language while yielding more concise and consistent feature descriptions. Their inherent structure affords new types of analyses, including quantifying feature complexity across layers, scaling automated interpretability from insights into individual features to model-wide patterns. Finally, in user studies, we find that semantic regexes help people build accurate mental models of LLM features.

cs.CL↗

arXiv2025

Visualization Vibes: The Socio-Indexical Function of Visualization Design

In contemporary information ecologies saturated with misinformation, disinformation, and a distrust of science itself, public data communication faces significant hurdles. Although visualization research has broadened criteria for effective design, governing paradigms privilege the accurate and efficient transmission of data. Drawing on theory from linguistic anthropology, we argue that such approaches-focused on encoding and decoding propositional content-cannot fully account for how people engage with visualizations and why particular visualizations might invite adversarial or receptive responses. In this paper, we present evidence that data visualizations communicate not only semantic, propositional meaning$\unicode{x2013}$meaning about data$\unicode{x2013}$but also social, indexical meaning$\unicode{x2013}$meaning beyond data. From a series of ethnographically-informed interviews, we document how readers make rich and varied assessments of a visualization's "vibes"$\unicode{x2013}$inferences about the social provenance of a visualization based on its design features. Furthermore, these social attributions have the power to influence reception, as readers' decisions about how to engage with a visualization concern not only content, or even aesthetic appeal, but also their sense of alignment or disalignment with the entities they imagine to be involved in its production and circulation. We argue these inferences hinge on a function of human sign systems that has thus far been little studied in data visualization: socio-indexicality, whereby the formal features (rather than the content) of communication evoke social contexts, identities, and characteristics. Demonstrating the presence and significance of this socio-indexical function in visualization, this paper offers both a conceptual foundation and practical intervention for troubleshooting breakdowns in public data communication.

cs.HC↗

arXiv2025

Quantifying Visualization Vibes: Measuring Socio-Indexicality at Scale

What impressions might readers form with visualizations that go beyond the data they encode? In this paper, we build on recent work that demonstrates the socio-indexical function of visualization, showing that visualizations communicate more than the data they explicitly encode. Bridging this with prior work examining public discourse about visualizations, we contribute an analytic framework for describing inferences about an artifact's social provenance. Via a series of attribution-elicitation surveys, we offer descriptive evidence that these social inferences: (1) can be studied asynchronously, (2) are not unique to a particular sociocultural group or a function of limited data literacy, and (3) may influence assessments of trust. Further, we demonstrate (4) how design features act in concert with the topic and underlying messages of an artifact's data to give rise to such 'beyond-data' readings. We conclude by discussing the design and research implications of inferences about social provenance, and why we believe broadening the scope of research on human factors in visualization to include sociocultural phenomena can yield actionable design recommendations to address urgent challenges in public data communication.

cs.HC↗

arXiv2025

Semantic Scaffolding: Augmenting Textual Structures with Domain-Specific Groupings for Accessible Data Exploration

Drawing connections between interesting groupings of data and their real-world meaning is an important, yet difficult, part of encountering a new dataset. A lay reader might see an interesting visual pattern in a chart but lack the domain expertise to explain its meaning. Or, a reader might be familiar with a real-world concept but struggle to express it in terms of a dataset's fields. In response, we developed semantic scaffolding, a technique for using domain-specific information from large language models (LLMs) to identify, explain, and formalize semantically meaningful data groupings. We present groupings in two ways: as semantic bins, which segment a field into domain-specific intervals and categories; and data highlights, which annotate subsets of data records with their real-world meaning. We demonstrate and evaluate this technique in Olli, an accessible visualization tool that exemplifies tensions around explicitly defining groupings while respecting the agency of readers to conduct independent data exploration. We conducted a study with 15 blind and low-vision (BLV) users and found that readers used semantic scaffolds to quickly understand the meaning of the data, but were often also critically aware of its influence on their interpretation.

cs.HC↗

arXiv2025

Tactile Vega-Lite: Rapidly Prototyping Tactile Charts with Smart Defaults

Tactile charts are essential for conveying data to blind and low vision (BLV) readers but are difficult for designers to construct. Non-expert designers face barriers to entry due to complex guidelines, while experts struggle with fragmented and time-consuming workflows that involve extensive customization. Inspired by formative interviews with expert tactile graphics designers, we created Tactile Vega-Lite (TVL): an extension of Vega-Lite that offers tactile-specific abstractions and synthesizes existing guidelines into a series of smart defaults. Predefined stylistic choices enable non-experts to produce guideline-compliant tactile charts quickly. Expert users can override defaults to tailor customizations for their intended audience. In a user study with 12 tactile graphics creators, we show that Tactile Vega-Lite enhances flexibility and consistency by automating tasks like adjusting spacing and translating braille while accelerating iterations through pre-defined textures and line styles. Through expert critique, we also learn more about tactile chart design best practices and design decisions.

cs.HC↗

arXiv2025

Pluto: Authoring Semantically Aligned Text and Charts for Data-Driven Communication

Textual content (including titles, annotations, and captions) plays a central role in helping readers understand a visualization by emphasizing, contextualizing, or summarizing the depicted data. Yet, existing visualization tools provide limited support for jointly authoring the two modalities of text and visuals such that both convey semantically-rich information and are cohesively integrated. In response, we introduce Pluto, a mixed-initiative authoring system that uses features of a chart's construction (e.g., visual encodings) as well as any textual descriptions a user may have drafted to make suggestions about the content and presentation of the two modalities. For instance, a user can begin to type out a description and interactively brush a region of interest in the chart, and Pluto will generate a relevant auto-completion of the sentence. Similarly, based on a written description, Pluto may suggest lifting a sentence out as an annotation or the visualization's title, or may suggest applying a data transformation (e.g., sort) to better align the two modalities. A preliminary user study revealed that Pluto's recommendations were particularly useful for bootstrapping the authoring process and helped identify different strategies participants adopt when jointly authoring text and charts. Based on study feedback, we discuss design implications for integrating interactive verification features between charts and text, offering control over text verbosity and tone, and enhancing the bidirectional flow in unified text and chart authoring tools.

cs.HC↗

arXiv2024

Toward Cultural Interpretability: A Linguistic Anthropological Framework for Describing and Evaluating Large Language Models (LLMs)

This article proposes a new integration of linguistic anthropology and machine learning (ML) around convergent interests in both the underpinnings of language and making language technologies more socially responsible. While linguistic anthropology focuses on interpreting the cultural basis for human language use, the ML field of interpretability is concerned with uncovering the patterns that Large Language Models (LLMs) learn from human verbal behavior. Through the analysis of a conversation between a human user and an LLM-powered chatbot, we demonstrate the theoretical feasibility of a new, conjoint field of inquiry, cultural interpretability (CI). By focusing attention on the communicative competence involved in the way human users and AI chatbots co-produce meaning in the articulatory interface of human-computer interaction, CI emphasizes how the dynamic relationship between language and culture makes contextually sensitive, open-ended conversation possible. We suggest that, by examining how LLMs internally "represent" relationships between language and culture, CI can: (1) provide insight into long-standing linguistic anthropological questions about the patterning of those relationships; and (2) aid model developers and interface designers in improving value alignment between language models and stylistically diverse speakers and culturally diverse speech communities. Our discussion proposes three critical research axes: relativity, variation, and indexicality.

cs.CY↗

arXiv2024

Charting EDA: Characterizing Interactive Visualization Use in Computational Notebooks with a Mixed-Methods Formalism

Interactive visualizations are powerful tools for Exploratory Data Analysis (EDA), but how do they affect the observations analysts make about their data? We conducted a qualitative experiment with 13 professional data scientists analyzing two datasets with Jupyter notebooks, collecting a rich dataset of interaction traces and think-aloud utterances. By qualitatively coding participant utterances, we introduce a formalism that describes EDA as a sequence of analysis states, where each state is comprised of either a representation an analyst constructs (e.g., the output of a data frame, an interactive visualization, etc.) or an observation the analyst makes (e.g., about missing data, the relationship between variables, etc.). By applying our formalism to our dataset, we identify that interactive visualizations, on average, lead to earlier and more complex insights about relationships between dataset attributes compared to static visualizations. Moreover, by calculating metrics such as revisit count and representational diversity, we uncover that some representations serve more as "planning aids" during EDA rather than tools strictly for hypothesis-answering. We show how these measures help identify other patterns of analysis behavior, such as the "80-20 rule", where a small subset of representations drove the majority of observations. Based on these findings, we offer design guidelines for interactive exploratory analysis tooling and reflect on future directions for studying the role that visualizations play in EDA.

cs.HC↗

arXiv2024

Abstraction Alignment: Comparing Model-Learned and Human-Encoded Conceptual Relationships

While interpretability methods identify a model's learned concepts, they overlook the relationships between concepts that make up its abstractions and inform its ability to generalize to new data. To assess whether models' have learned human-aligned abstractions, we introduce abstraction alignment, a methodology to compare model behavior against formal human knowledge. Abstraction alignment externalizes domain-specific human knowledge as an abstraction graph, a set of pertinent concepts spanning levels of abstraction. Using the abstraction graph as a ground truth, abstraction alignment measures the alignment of a model's behavior by determining how much of its uncertainty is accounted for by the human abstractions. By aggregating abstraction alignment across entire datasets, users can test alignment hypotheses, such as which human concepts the model has learned and where misalignments recur. In evaluations with experts, abstraction alignment differentiates seemingly similar errors, improves the verbosity of existing model-quality metrics, and uncovers improvements to current human abstractions.

cs.LG↗

arXiv2024

Umwelt: Accessible Structured Editing of Multimodal Data Representations

We present Umwelt, an authoring environment for interactive multimodal data representations. In contrast to prior approaches, which center the visual modality, Umwelt treats visualization, sonification, and textual description as coequal representations: they are all derived from a shared abstract data model, such that no modality is prioritized over the others. To simplify specification, Umwelt evaluates a set of heuristics to generate default multimodal representations that express a dataset's functional relationships. To support smoothly moving between representations, Umwelt maintains a shared query predicated that is reified across all modalities -- for instance, navigating the textual description also highlights the visualization and filters the sonification. In a study with 5 blind / low-vision expert users, we found that Umwelt's multimodal representations afforded complementary overview and detailed perspectives on a dataset, allowing participants to fluidly shift between task- and representation-oriented ways of thinking.

cs.HC↗

arXiv2023

DiffusionWorldViewer: Exposing and Broadening the Worldview Reflected by Generative Text-to-Image Models

Generative text-to-image (TTI) models produce high-quality images from short textual descriptions and are widely used in academic and creative domains. Like humans, TTI models have a worldview, a conception of the world learned from their training data and task that influences the images they generate for a given prompt. However, the worldviews of TTI models are often hidden from users, making it challenging for users to build intuition about TTI outputs, and they are often misaligned with users' worldviews, resulting in output images that do not match user expectations. In response, we introduce DiffusionWorldViewer, an interactive interface that exposes a TTI model's worldview across output demographics and provides editing tools for aligning output images with user perspectives. In a user study with 18 diverse TTI users, we find that DiffusionWorldViewer helps users represent their varied viewpoints in generated images and challenge the limited worldview reflected in current TTI models.

cs.LG↗

arXiv2023

Heuristics for Supporting Cooperative Dashboard Design

Dashboards are no longer mere static displays of metrics; through functionality such as interaction and storytelling, they have evolved to support analytic and communicative goals like monitoring and reporting. Existing dashboard design guidelines, however, are often unable to account for this expanded scope as they largely focus on best practices for visual design. In contrast, we frame dashboard design as facilitating an analytical conversation: a cooperative, interactive experience where a user may interact with, reason about, or freely query the underlying data. By drawing on established principles of conversational flow and communication, we define the concept of a cooperative dashboard as one that enables a fruitful and productive analytical conversation, and derive a set of 39 dashboard design heuristics to support effective analytical conversations. To assess the utility of this framing, we asked 52 computer science and engineering graduate students to apply our heuristics to critique and design dashboards as part of an ungraded, opt-in homework assignment. Feedback from participants demonstrates that our heuristics surface new reasons dashboards may fail, and encourage a more fluid, supportive, and responsive style of dashboard design. Our approach suggests several compelling directions for future work, including dashboard authoring tools that better anticipate conversational turn-taking, repair, and refinement and extending cooperative principles to other analytical workflows.

cs.HC↗

arXiv2023

"Customization is Key": Reconfigurable Content Tokens for Accessible Data Visualizations

Customization is crucial for making visualizations accessible to blind and low-vision (BLV) people with widely-varying needs. But what makes for usable or useful customization? We identify four design goals for how BLV people should be able to customize screen-reader-accessible visualizations: presence, or what content is included; verbosity, or how concisely content is presented; ordering, or how content is sequenced; and, duration, or how long customizations are active. To meet these goals, we model a customization as a sequence of content tokens, each with a set of adjustable properties. We instantiate our model by extending Olli, an open-source accessible visualization toolkit, with a settings menu and command box for persistent and ephemeral customization respectively. Through a study with 13 BLV participants, we find that customization increases the ease of identifying and remembering information. However, customization also introduces additional complexity, making it more helpful for users familiar with similar tools.

cs.HC↗

arXiv2023

Bluefish: Composing Diagrams with Declarative Relations

Diagrams are essential tools for problem-solving and communication as they externalize conceptual structures using spatial relationships. But when picking a diagramming framework, users are faced with a dilemma. They can either use a highly expressive but low-level toolkit, whose API does not match their domain-specific concepts, or select a high-level typology, which offers a recognizable vocabulary but supports a limited range of diagrams. To address this gap, we introduce Bluefish: a diagramming framework inspired by component-based user interface (UI) libraries. Bluefish lets users create diagrams using relations: declarative, composable, and extensible diagram fragments that relax the concept of a UI component. Unlike a component, a relation does not have sole ownership over its children nor does it need to fully specify their layout. To render diagrams, Bluefish extends a traditional tree-based scenegraph to a compound graph that captures both hierarchical and adjacent relationships between nodes. To evaluate our system, we construct a diverse example gallery covering many domains including mathematics, physics, computer science, and even cooking. We show that Bluefish's relations are effective declarative primitives for diagrams. Bluefish is open source, and we aim to shape it into both a usable tool and a research platform.

cs.GR↗

arXiv2023

VisText: A Benchmark for Semantically Rich Chart Captioning

Captions that describe or explain charts help improve recall and comprehension of the depicted data and provide a more accessible medium for people with visual disabilities. However, current approaches for automatically generating such captions struggle to articulate the perceptual or cognitive features that are the hallmark of charts (e.g., complex trends and patterns). In response, we introduce VisText: a dataset of 12,441 pairs of charts and captions that describe the charts' construction, report key statistics, and identify perceptual and cognitive phenomena. In VisText, a chart is available as three representations: a rasterized image, a backing data table, and a scene graph -- a hierarchical representation of a chart's visual elements akin to a web page's Document Object Model (DOM). To evaluate the impact of VisText, we fine-tune state-of-the-art language models on our chart captioning task and apply prefix-tuning to produce captions that vary the semantic content they convey. Our models generate coherent, semantically rich captions and perform on par with state-of-the-art chart captioning models across machine translation and text generation metrics. Through qualitative analysis, we identify six broad categories of errors that our models make that can inform future work.

cs.CV↗

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↗