Visualization dashboards are regularly used for data exploration and analysis, but their complex interactions and interlinked views often require time-consuming onboarding sessions from dashboard authors. Preparing these onboarding materials is labor-intensive and requires manual updates when dashboards change. Recent advances in multimodal interaction powered by large language models (LLMs) provide ways to support self-guided onboarding. We present DIANA (Dashboard Interactive Assistant for Navigation and Analysis), a multimodal dashboard assistant that helps users for navigation and guided analysis through chat, audio, and mouse-based interactions. Users can choose any interaction modality or a combination of them to onboard themselves on the dashboard. Each modality highlights relevant dashboard features to support user orientation. Unlike typical LLM systems that rely solely on text-based chat, DIANA combines multiple modalities to provide explanations directly in the dashboard interface. We conducted a comparative qualitative user study to understand the use of different modalities for different types of onboarding tasks and their complexities.
Large Language Models (LLMs) are transforming Conversational Visual Analytics (CVA) by enabling data analysis through natural language. However, evaluating LLMs for CVA remains a challenge: requiring programming expertise, overlooking real-world complexity, and lacking interpretable metrics for multi-format (visualizations and text) outputs. Through interviews with 22 CVA developers and 16 end-users, we identified use-cases, evaluation criteria and workflows. We present Lexara, a user-centered evaluation toolkit for CVA that operationalizes these insights into: (i) test-cases spanning real-world scenarios; (ii) interpretable metrics covering visualization quality (data fidelity, semantic alignment, functional correctness, design clarity) and language quality (factual grounding, analytical reasoning, conversational coherence) using rule-based and LLM-as-a-judge methods; and (iii) an interactive toolkit enabling experimental setup and multi-format and multi-level exploration of results without programming expertise. We conducted a two-week diary study with six CVA developers, drawn from our initial cohort of 22. Their feedback demonstrated Lexara's effectiveness for guiding appropriate model and prompt selection.
Design feedback helps practitioners improve their artifacts while also fostering reflection and design reasoning. Large Language Models (LLMs) such as ChatGPT can support design work, but often provide generic, one-off suggestions that limit reflective engagement. We investigate how to guide LLMs to act as design mentors by applying the Cognitive Apprenticeship Model, which emphasizes demonstrating reasoning through six methods: modeling, coaching, scaffolding, articulation, reflection, and exploration. We operationalize these instructional methods through structured prompting and evaluate them in a within-subjects study with data visualization practitioners. Participants interacted with both a baseline LLM and an instructional LLM designed with cognitive apprenticeship prompts. We further conducted surveys, interviews, and conversational log analyses to evaluate experiences across conditions. Our findings show that cognitively informed prompts elicit deeper design reasoning and more reflective feedback exchanges, though the baseline is sometimes preferred depending on task types or experience levels. We distill design considerations for AI-assisted feedback systems that foster reflective practice.
Hyperparameter optimization (HPO) is a long-running process that can span hours or even days. While recent Human-in-the-Loop HPO systems enable monitoring and steering of the process, they are typically designed for desktop environments, which limits their effectiveness in managing prolonged experiments in practice. To address these limitations, we present HyPockeTuner, an interactive mobile system that enables users to monitor, steer, and reflect on HPO experiments anytime, anywhere from smartphones. Its mobile-tailored interface supports tracking experiment history and visualizing the relationship between user interventions and performance changes. HyPockeTuner also employs a notification workflow that alerts users to important events, reducing the burden of constant monitoring while enabling timely interventions. In a pilot study, we validated that users could readily identify critical events, such as performance improvements and intervention points, through our visualization. Furthermore, two five-day deployment studies with follow-up reflection sessions demonstrated that users could integrate experiment management into their daily routines and reflect on past decisions, generating insights for future improvement.
The Transformer architecture underpins modern large language models powering state-of-the-art text generation and AI applications. However, its complexity makes it difficult for non-experts to learn. Existing resources often lack interactivity, rely on static descriptions of simplified architectures, or fail to reflect models’ behavior with real data. To address this gap, we introduce Transformer Explainer, an interactive visualization tool for non-experts to learn Transformers. The tool integrates an overview illustrating the Transformer's data flow with on-demand explanations that gradually reveal mathematical details. Smooth transitions across abstraction levels highlight the interplay between high-level structures and low-level operations. Running a live GPT-2 instance directly in the browser, Transformer Explainer empowers learners to experiment with custom input and hyperparameters without setup, observing next-token predictions in real time. A 90-participant user study showed that our tool offered significant advantages in improving user understanding and engagement. Transformer Explainer has attracted over 490,000 users.
Chart data extraction, which reverse-engineers data tables from chart images, is essential for reproducibility, analysis, retrieval, and redesign. Existing interactive tools are reliable but tedious, and mixed-initiative systems, while more efficient, lack generalizability. Recent multimodal large language models (MLLMs) offer a unified interface for chart interpretation, yet their ability to extract accurate data tables, especially without visible labels, remains unclear. We build a benchmark featuring diverse real-world charts without data labels to evaluate this capability. Results show that, while current MLLMs reliably reconstruct table structures, they struggle with precise value recovery. To address this, we revisit chart data extraction from a human-centered perspective and argue that extraction should follow a progressive learning process similar to how people read charts. Our training framework substantially improves numerical accuracy, achieving state-of-the-art performance with a 7B-parameter model. A user study further shows that our model effectively supports mixed-initiative workflows for reliable chart data extraction.