Pith. sign in

REVIEW 4 major objections 7 minor 73 references

VizPilot: Automated Onboarding for SVG-based Composite Visualizations using Multimodal LLMs

T0 review · 4 major / 7 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read VizPilot claims it can automatically generate interactive guided tours for complex multi-chart visualizations straight from the SVG code and a screenshot, with no manual authoring.

desk verdict VizPilot is a solid systems contribution with a clever mapping validator and a fair evaluation, but the 'no manual authoring' claim is not supported because the developer description is never ablated. read the letter →

arxiv 2607.27938 v1 pith:YVB5ZHVD submitted 2026-07-30 cs.HC

Nishaanthini Gnanavel , Yong Wang This is my paper · ORCID
classification cs.HC
keywords compositevisualizationsvisualizationonboardingmultimodallargelanguagemodelsSVGsemanticmappinghierarchicalselectorinferencescrollytellingcognitiveloadbrowserextension
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

VizPilot claims that interactive onboarding for complex composite visualizations can be generated automatically, without manual authoring, from the raw artifacts a developer already has: the SVG code, a rendered bitmap screenshot, a brief description, and optionally the interaction source code. The system decomposes the visualization into visual components, extracts structured knowledge about each one, and maps every explanation to the exact SVG elements that should be highlighted. If this works, developers no longer need to hand-write tour scripts or bind events by hand, and first-time viewers of unfamiliar multi-chart displays can learn them through guided scrolling and free-form exploration instead of staring at static text. The paper supports the claim with a gallery of 18 visualizations spanning all four composite design patterns and a user study in which 16 participants answered questions faster and with lower reported mental demand using VizPilot than with text-only descriptions.

What carries the argument

The load-bearing object is the Composite Visualization Analyzer, a two-stage MLLM pipeline. Semantic Inference decomposes the composite visualization into visual components and emits atomic explanation units — one sentence, one component, one knowledge category — using a five-category taxonomy drawn from onboarding literature. Semantic Mapping then performs Hierarchical Selector Inference, prompting the model to write CSS-like selector paths from parent containers down to target marks rather than raw IDs, and passes every candidate selector through a three-check programmatic validation: DOM validity, component containment, and layout consistency. This combination converts probabilistic model output into deterministic, verifiable highlighting targets.

What would settle it

Run VizPilot on a composite visualization whose SVG has been flattened into a single path with no grouped containers: if the system cannot produce any valid selectors or onboarding steps, that confirms the central raw-artifact claim is limited to structurally grouped SVG inputs. A complementary check is to apply it to a canvas-rendered visualization; the paper's own scope statement predicts component extraction accuracy drops.

Watch

Extended reading notes

Core claim

The paper's central discovery is that a multimodal LLM, constrained by a structured two-stage pipeline, can reverse-engineer the semantic structure of a composite SVG visualization and turn it into a mapped, interactive onboarding experience. The first stage, Semantic Inference, decomposes the chart into visual components and generates atomic explanation units under a five-category knowledge taxonomy (structure, data, encoding, analytical, interaction). The second stage, Semantic Mapping, anchors each unit to DOM nodes via hierarchical selector inference, then verifies the selectors programmatically. The claimed result is end-to-end onboarding generation with a mean end-to-end success rate of 0.898 across 18 gallery visualizations, and a user study showing significantly faster task completion (194 seconds vs. 263 seconds), lower NASA-TLX mental demand, effort, and frustration, and higher PSSUQ ratings, with task accuracy at ceiling in both conditions.

Load-bearing premise

The pipeline works only when the target visualization exposes a hierarchical, DOM-queryable SVG structure; canvas-rendered charts and heavily flattened SVGs are outside its scope and would break the semantic mapping stage.

Editorial extensions

If this is right

  • Visualization developers can publish onboarding for a composite chart by providing a brief description and optional interaction code; the rest of the tour is generated and exported as a JSON file plus a lightweight runtime script.
  • End users of unfamiliar composite visualizations get two complementary paths — guided narrative scrollytelling with synchronized highlighting, and free exploration with hover inspectors and a context-constrained assistant — which the study ties to reduced split-attention effort.
  • The approach generalizes across the four composite design patterns (juxtaposition, overloading, superimposition, nesting), with exact mapping accuracy above 0.89 and end-to-end success above 0.80 in the gallery evaluation.
  • Because all mappings are expressed as validated SVG selectors, the generated onboarding can be deployed without modifying the visualization's underlying source code.
  • Interaction knowledge, when interaction source code is available, shifts from guesswork to verified action-consequence descriptions of hover, click, brush, and linked updates.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The same mapped semantic representation could power accessibility features beyond onboarding, such as screen-reader narration, keyboard-navigable tours, and automated alternative text, since each explanation is already anchored to precise DOM elements.
  • If the mapping mechanism generalizes, it could turn the pipeline into a general evaluation harness for whether multimodal models truly understand chart structure, by scoring decomposition and grounding quality on any SVG-based visualization.
  • A testable extension is to apply the decomposition to dashboards composed of multiple SVG containers; the paper currently assumes a single SVG container, so multi-view dashboards would be the natural next boundary.
  • The flattened-SVG limitation suggests a targeted stress test: feed the system an SVG where all shapes are merged into one path, and measure how much decomposition recall and mapping accuracy drop; the paper's own scope statement predicts a sharp drop.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 7 minor

Summary. VizPilot is a browser extension that generates interactive onboarding experiences for SVG-based composite visualizations. It takes as input the extracted SVG code, a rendered bitmap image, a developer-provided brief description, and optionally the interaction source code. A two-stage Composite Visualization Analyzer first performs Semantic Inference to decompose the visualization into visual components and produce atomic explanation units organized by a five-category knowledge taxonomy, then Semantic Mapping anchors these units to specific SVG elements via hierarchical selector inference with programmatic validation. The Onboarding Interface provides narrative scrollytelling and free exploration. The paper evaluates the approach with a comparative input-modality analysis, a gallery of 18 visualizations scored by LLM-as-judge, a usage scenario with expert feedback, and a 16-participant user study contrasting VizPilot against a text baseline.

Significance. If the central claims hold, VizPilot addresses a real gap: prior visualization onboarding approaches are platform-dependent, require substantial manual authoring, and mostly target simple charts or dashboards, whereas VizPilot targets customized composite SVG visualizations in a browser context. The two-stage decomposition with hierarchical selector inference and programmatic DOM validation is a sensible design that combines semantic reasoning with deterministic grounding. The paper contributes a working implementation, a public gallery, a multimodal ablation of SVG/image/interaction-code inputs, and a user study with both quantitative and qualitative data. These are concrete strengths. However, the significance is tempered by three issues: the developer-provided description is never ablated, so the claim of reverse-engineering 'directly from raw visualization artifacts' is not fully established; the gallery evaluation depends on author-defined ground truth and LLM-as-judge with manual escalation; and the user study has a ceiling effect on accuracy and a single marginal objective performance result.

major comments (4)
  1. [§1, §3, Appendix A.1] The central claim of automated reverse-engineering 'directly from raw visualization artifacts' is not adequately supported because the developer-provided description is a primary input to Semantic Inference and is never ablated. The Semantic Inference prompt in Appendix A.1 lists the 'user-authored description' as one of the four inputs, and the modality analysis in §7.1 varies SVG, image, and interaction source code but keeps the description fixed in all conditions. Consequently, component names, roles, relationships, and reading order may be largely supplied by the description rather than reconstructed from the SVG. I request an ablation that runs the pipeline without the description and reports decomposition, mapping, and explanation-quality metrics, and a revised claim that reflects the actual input requirements.
  2. [§7.3, Table 3] The gallery evaluation relies on author-defined ground-truth component lists and an LLM-as-judge protocol with a confidence threshold of 0.7; cases below the threshold (5 of 18) are escalated to manual review by the authors. Because the ground truth and rubric are derived from the same component definition that the pipeline implements, the reported E2E success rate of 0.898 may be optimistic. Please report inter-judge agreement among the three judge models, results computed over the full set excluding manually reviewed cases, and a pre-registered or independently verified random sample of the manual-review decisions.
  3. [§7.4.1] The user study's only objective performance benefit is task completion time (p=0.039) from a 16-participant, two-visualization design, and task accuracy shows a ceiling effect (baseline 3.62/4 vs. VizPilot 3.75/4). Although the NASA-TLX reductions are large and consistent, the paper should report effect sizes and confidence intervals for the time and workload results, and should discuss whether the p=0.039 task-time result would remain significant after correcting for the number of Wilcoxon tests performed. As written, the statement that 'VizPilot accelerates insight extraction' rests on a single marginal p-value.
  4. [§7.2, Appendix A.7] The claimed reduction in developer authoring effort is supported only by a single usage scenario and an interview with two experts. No quantitative comparison against a manual authoring baseline (e.g., authoring time, number of editing operations, or interaction-code adjustments) is provided. Without such a comparison, the 'no manual authoring effort' claim remains a design assertion rather than an evaluated outcome. Adding a small comparative study or at least a structured time measurement for the authoring workflow would materially strengthen the paper.
minor comments (7)
  1. [Abstract and §1] The abstract and §1 say VizPilot 'requires no manual authoring effort,' while the abstract also says it 'requires only a brief visualization description and optional interaction source code.' These statements are in tension; please rephrase to avoid an overclaim, e.g., 'requires a brief description but no step-by-step authoring.'
  2. [§8.1] The limitations regarding canvas-rendered visualizations and heavily flattened SVGs are stated only in the Discussion. Since they directly qualify the platform-agnostic framing, they should be mentioned prominently in the Introduction or the system overview.
  3. [§7.4] The procedure section says 'three phrases' where it means 'three phases'; please fix this typo.
  4. [Table 5] The row labeled 'Multiple Time SeriesMultiple Time Series' contains a duplicated phrase; the label should be 'Multiple Time Series.'
  5. [§8.2] The phrase 'maintaining user anatomy' appears to be a typo; it should likely be 'user agency' or 'user autonomy.'
  6. [Figures 6 and 7] The figures mark several comparisons with '*** (p<0.001)' but the exact p-values are not reported in the text for each item; please list them in the figure captions or in the results section.
  7. [§6 and §7.3] The implementation uses GPT-5 for inference while the gallery judges include GPT-5.5, Gemini 3.1 Pro Preview, and Claude 3.5 Sonnet; please clarify whether the inference model is the same across all evaluations and whether the judge models are distinct from the generation model, as this affects reproducibility and the interpretation of the LLM-as-judge results.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the pipeline's outputs are not definitionally contained in its inputs, and the developer-provided description, while not ablated, is an input rather than a fitted parameter.

full rationale

I walked the VizPilot derivation chain from multimodal inputs (SVG code, bitmap, brief description, optional interaction source code) to the intermediate semantic representation, then to mapped SVG selectors and the onboarding interface. No equation-level or construction-level reduction is present: the paper does not define any output in terms of its own inputs, does not fit a parameter to a subset of data and then predict a closely related quantity, and does not invoke a uniqueness theorem to forbid alternatives. The user-authored description is a genuine input, not a renamed prediction; the paper is transparent that the workflow requires it, and the claim being made is that onboarding steps need not be manually authored, not that the system runs with zero textual input. The lack of an ablation removing the description is an evaluation gap and a threat to the strength of the claim, but it is not a circular derivation because the paper never claims the description is absent or that the output is derived from it by construction. The decomposition ground truth is defined using the paper's own visual-component criterion, which weakens the external validity of the decomposition F1 scores as an independent benchmark; however, the system must still apply that criterion to novel SVG structures, so the evaluation is a consistency check rather than a formal self-fulfilling reduction. LLM-as-judge evaluations receive the pipeline output and a rubric, which is a self-assessment protocol, not a circular derivation of the central mechanism. Self-citations in the paper (e.g., PrettiSmart, PonziLens, QuantumEyes, Diffseer) appear as gallery examples and related-work context, not as load-bearing evidence for the reverse-engineering approach. The paper also states honest scope limitations in Section 8.1 (structurally grouped, DOM-accessible SVGs assumed; canvas and heavily flattened SVGs excluded) and Section 8.3 (single SVG container, MLLM context limits, novelty effects), which further supports that the contribution is an engineering pipeline with empirical evaluation rather than a derivation whose conclusion is already contained in its premises. Overall, no load-bearing circular step was found.

Assumptions & free parameters 1 free parameters · 5 assumptions · 0 invented entities

No new physical or formal entities are postulated. 'Atomic explanation units' and 'hierarchical selector inference' are methodological constructs, not entities with independent falsifiable handles. The central dependencies are the MLLM's competence, the SVG structure assumption, and the evaluation protocol's validity.

free parameters (1)
  • LLM-as-judge escalation threshold = 0.7
    Hand-chosen confidence threshold in Appendix A.6: cases with any criterion below 0.7 were escalated to manual review by the authors. This threshold affects the reported gallery accuracy numbers; it is an evaluation parameter, not a pipeline parameter.
assumptions (5)
  • domain assumption Structurally grouped, DOM-accessible SVG is available for the target visualization
    Stated in Section 8.1; Semantic Mapping in Section 4.2 depends on a queryable SVG DOM to resolve and verify selectors.
  • domain assumption The MLLM (GPT-5) can perform visual component decomposition, knowledge extraction, and hierarchical selector inference from the constrained prompts
    Sections 4.1, 4.2, and Appendices A.1 and A.3; the entire semantic layer rests on empirical MLLM competence.
  • domain assumption The five-category knowledge taxonomy (structure, data, encoding, analytical, interaction) is appropriate and sufficient for onboarding explanations
    Section 4.1 and Table 1; borrowed from onboarding literature [41] and operationalized here.
  • domain assumption LLM-as-judge evaluation with the provided rubric is a valid proxy for ground-truth quality of decomposition, mapping, and coherence
    Section 7.3 and Appendix A.6; 5 of 18 visualizations were escalated to manual author review, indicating the judges were not fully trusted.
  • ad hoc to paper The component definition (a visual structure qualifies iff it independently supports extraction of at least one distinct form of visualization knowledge) yields unambiguous ground truth
    Defined in Section 4.1 and Appendix A.5, used to build ground-truth lists in Appendix A.6, creating a partial self-reference in evaluation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of VizPilot: Automated Onboarding for SVG-based Composite Visualizations using Multimodal LLMs." pith.science (2026). https://pith.science/paper/YVB5ZHVD

@misc{pith2026260727938,
  author       = {Pith},
  title        = {Pith review of: VizPilot: Automated Onboarding for SVG-based Composite Visualizations using Multimodal LLMs},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YVB5ZHVD}},
  note         = {Machine review of arXiv:2607.27938}
}
read the original abstract

Composite visualizations integrate multiple visualizations to represent complex datasets effectively, but their intrinsic composite designs often impose a high initial cognitive load on novice users. Existing visualization onboarding approaches are typically platform-dependent, require substantial manual authoring effort, and struggle with the structural complexity of composite visualizations, limiting their general applicability. We present VizPilot, an automated visualization onboarding approach that reverse-engineers composite visualization structure to generate interactive onboarding experiences directly from raw visualization artifacts. VizPilot consists of two modules: a Composite Visualization Analyzer and an Onboarding Interface. Leveraging Multimodal Large Language Models (MLLMs), the Analyzer employs a two-stage pipeline that decomposes a visualization into visual components, extracts structured explanations, and maps them to precise SVG elements for reliable highlighting and interaction. Implemented as a browser extension, VizPilot requires only a brief visualization description and optional interaction source code from the visualization developer to automatically generate onboarding content. The Onboarding Interface supports both guided narrative scrollytelling and free exploration, enabling users to learn visualization components progressively or on demand. We evaluate VizPilot through a comparative analysis of different input modalities, a usage scenario demonstrating reduced authoring effort, and a user study assessing its impact on users' cognitive load. The results demonstrate that VizPilot effectively automates the authoring of onboarding experiences while improving the usability and accessibility of composite visualizations.

Figures

Figures reproduced from arXiv: 2607.27938 by the authors.

Figure 1
Figure 1. VizPilot Overview. (A) VizPilot consumes multimodal inputs including SVG code, bitmap images, developer-provided description and optional interaction source code (example chart from Highcharts [19]). (B) The Composite Visualization Analyzer processes these via two stages: (B1) Semantic Inference decomposes the chart to extract knowledge and generate explanation text, producing (B3) intermediate semantic representati… view at source ↗
Figure 2
Figure 2. The Semantic Mapping Stage. VizPilot maps each semantic unit from the (A) Intermediate Semantic Representation into precise SVG elements. (B) An MLLM proposes candidate DOM selectors through a hierarchical selector inference for both component-level and encoding￾level targets, which are then (C) verified against a multi-level selector validation. The resulting (D) mapped explanation units enable synchro￾nized visual… view at source ↗
Figure 4
Figure 4. Usage Scenario: Authoring Workflow for PrettiSmart Composite Visualization [49]. (A) The developer, Alex, provides the interaction source code and a brief description of the visualization. (B) The system automatically generates a mapped onboarding sequence, which Alex reviews for accuracy. (C) Alex refines a specific explanation unit via the edit option. (D) Finally, Alex exports the onboarding content to deploy the… view at source ↗
Figures from the paper (4 more)
Figure 5
Figure 5. Figure 5: Gallery (A) The VizPilot gallery demonstrates generalizability across diverse composite visualizations. A representative onboarding sequence illustrates our narrative scrollytelling mechanism: as the user progresses through (B–D) ordered explanation units, VizPilot dyn…
Figure 6
Figure 6. Figure 6: PSSUQ Results. Results for selected PSSUQ items comparing the text baseline and VizPilot. VizPilot demonstrates a significant positive shift across both Information Quality Q10–Q12 and System Usefulness (Q3, Q5). Brackets with stars denote statistical significance base…
Figure 7
Figure 7. Figure 7: NASA-TLX Workload Assessment. Box plots comparing the perceived workload across selected dimensions for the baseline (shown in green color) and VizPilot (shown in blue color). Except for Perfor￾mance, lower scores indicate a more favorable, lower-friction experience. V…
Figure 8
Figure 8. Figure 8: Component decomposition of a composite visualization from [PITH_FULL_IMAGE:figures/full_fig_p014_8.png]

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

73 extracted references · 31 canonical work pages

  1. [1]

    M. Z. I. Alam, S. Islam, and E. Hoque. Seechart: Enabling accessible vi- sualizations through interactive natural language interface for people with visual impairments. InProceedings of the 28th International Conference on Intelligent User Interfaces, IUI ’23, 19 pages, p. 46–64. Association for Computing Machinery, New York, NY , USA, 2023. doi: 10.1145/...

    doi:10.1145/3581641 2023

  2. [2]

    Apache ECharts Examples: Candlestick on Mobile

    Apache Software Foundation. Apache ECharts Examples: Candlestick on Mobile. https://echarts.apache.org/examples/en/editor. html?c=candlestick-touch. [Online].Accessed: 2026-06-24. 16

    work page 2026

  3. [3]

    Apache ECharts Examples: Shanghai In- dex

    Apache Software Foundation. Apache ECharts Examples: Shanghai In- dex. https://echarts.apache.org/examples/en/editor.html? c=candlestick-sh. [Online].Accessed: 2026-04-01. 6, 7, 15, 16

    work page 2026

  4. [5]

    Bostock, V

    M. Bostock, V . Ogievetsky, and J. Heer. D3 data-driven documents.IEEE Transactions on Visualization and Computer Graphics, 17(12):2301–2309, 9 pages, Dec. 2011. doi: 10.1109/TVCG.2011.185 2

    doi:10.1109/tvcg.2011.185 2011

  5. [6]

    C. Chen, H. K. Bako, P. Yu, J. Hooker, J. Joyal, S. C. Wang, S. Kim, J. Wu, A. Ding, L. Sandeep, A. Chen, C. Sinha, and Z. Liu. Visanatomy: An svg chart corpus with fine-grained semantic labels.IEEE Transactions on Visualization and Computer Graphics, 32(1):560–570, 2026. doi: 10. 1109/TVCG.2025.3634263 2

    arXiv 2026

  6. [7]

    C. Chen, B. Lee, Y . Wang, Y . Chang, and Z. Liu. Mystique: Deconstructing svg charts for layout reuse.IEEE Transactions on Visualization and Computer Graphics, 30(1):447–457, 2024. doi: 10.1109/TVCG.2023. 3327354 2

    doi:10.1109/tvcg.2023 2024

  7. [8]

    A. K. Das, M. Tarun, and K. Mueller. Charts-of-thought: Enhancing llm visualization literacy through structured data extraction.IEEE Transac- tions on Visualization and Computer Graphics, 32(1):1427–1437, 2026. doi: 10.1109/TVCG.2025.3634813 3

    arXiv 2026

  8. [9]

    D. Deng, W. Cui, X. Meng, M. Xu, Y . Liao, H. Zhang, and Y . Wu. Revis- iting the design patterns of composite visualizations.IEEE Transactions on Visualization and Computer Graphics, 29(12):5406–5421, 2023. doi: 10.1109/TVCG.2022.3213565 1

    arXiv 2023

Show all 73 references
  1. [10]

    Dhanoa, A

    V . Dhanoa, A. Hinterreiter, V . Fediuk, N. Elmqvist, E. Gröller, and M. Streit. D-tour: Semi-automatic generation of interactive guided tours for visualization dashboard onboarding.IEEE Transactions on Visualiza- tion and Computer Graphics, 31(1):721–731, 2025. doi: 10.1109/T...

    2025

  2. [11]

    Dhanoa, G

    V . Dhanoa, G. M. León, E. Hoggan, E. Gröller, M. Streit, and N. Elmqvist. Hey dashboard!: Supporting voice, text, and pointing modalities in dash- board onboarding.arXiv preprint arXiv:2510.12386, 2025. 2

    2025

  3. [12]

    Dhanoa, C

    V . Dhanoa, C. Walchshofer, A. Hinterreiter, H. Stitz, E. Groeller, and M. Streit. A process model for dashboard onboarding.Computer Graphics Forum, 41(3):501–513, 2022. doi: 10.1111/cgf.14558 2

    2022 doi

  4. [13]

    X. L. Do, M. Hassanpour, A. Masry, P. Kavehzadeh, E. Hoque, and S. Joty. Do llms work on charts? designing few-shot prompts for chart question answering and summarization.arXiv preprint arXiv:2312.10610, 2023. 3

    2023 arXiv

  5. [14]

    A. Dong, W. Zeng, X. Chen, and Z. Cheng. Vistory: Interactive storyboard for exploring visual information in scientific publications. InProceedings of the 12th International Symposium on Visual Information Communica- tion and Interaction, VINCI ’19, article no. 12, 8 pages. As...

    2019 doi

  6. [16]

    J. Gu, X. Jiang, Z. Shi, H. Tan, X. Zhai, C. Xu, W. Li, Y . Shen, S. Ma, H. Liu, S. Wang, K. Zhang, Z. Lin, B. Zhang, L. Ni, W. Gao, Y . Wang, and J. Guo. A survey on llm-as-a-judge.The Innovation, 7(6):101253,

  7. [17]

    Y . Han, C. Zhang, X. Chen, X. Yang, Z. Wang, G. Yu, B. Fu, and H. Zhang. Chartllama: A multimodal llm for chart understanding and generation. arXiv preprint arXiv:2311.16483, 2023. 3

    2023 arXiv

  8. [18]

    S. G. Hart and L. E. Staveland. Development of nasa-tlx (task load index): Results of empirical and theoretical research. In P. A. Hancock and N. Meshkati, eds.,Human Mental Workload, vol. 52 ofAdvances in Psychology, pp. 139–183. North-Holland, 1988. doi: 10.1016/S0166-4115 (...

    1988 doi

  9. [19]

    Highcharts demo: Combo multi-axes chart

    Highsoft AS. Highcharts demo: Combo multi-axes chart. https:// www.highcharts.com/demo/highcharts/combo-multi-axes . [On- line].Accessed: 2026-03-30. 1

    2026

  10. [20]

    J. Hong, C. Seto, A. Fan, and R. Maciejewski. Do llms have visualization literacy? an evaluation on modified visualizations to test generalization in data interpretation.IEEE Transactions on Visualization and Computer Graphics, 31(10):7004–7018, 2025. doi: 10.1109/TVCG.2025.3536358 3

    2025

  11. [21]

    Hoque and N

    N. Hoque and N. Sultanum. Dashguide: Authoring interactive dashboard tours for guiding dashboard users.Computer Graphics Forum, 44, 05

  12. [22]

    Javed and N

    W. Javed and N. Elmqvist. Exploring the design space of composite visualization. In2012 IEEE Pacific Visualization Symposium, pp. 1–8,

  13. [23]

    S. R. Khan, V . Chandak, and S. Mukherjea. Evaluating llms for visualiza- tion generation and understanding.Discover Data, 3(1):15, 2025. doi: 10. 1007/s44248-025-00036-4 3

    2025

  14. [24]

    Kosch, R

    T. Kosch, R. Welsch, L. Chuang, and A. Schmidt. The placebo effect of artificial intelligence in human–computer interaction.ACM Trans. Comput.-Hum. Interact., 29(6), article no. 56, 32 pages, Jan. 2023. doi: 10 .1145/3529225 8, 9

    2023

  15. [25]

    C. Lai, Z. Lin, R. Jiang, Y . Han, C. Liu, and X. Yuan. Automatic annotation synchronizing with textual description for visualization. InProceedings of the 2020 CHI Conference on Human Factors in Computing Systems, CHI ’20, 13 pages, p. 1–13. Association for Computing Machiner...

    2020

  16. [26]

    Lee, S.-H

    S. Lee, S.-H. Kim, and B. C. Kwon. Vlat: Development of a visualization literacy assessment test.IEEE Transactions on Visualization and Computer Graphics, 23(1):551–560, 2017. doi: 10.1109/TVCG.2016.2598920 3

    2017

  17. [27]

    J. R. Lewis. Ibm computer usability satisfaction questionnaires: Psy- chometric evaluation and instructions for use.International Jour- nal of Human–Computer Interaction, 7(1):57–78, 1995. doi: 10.1080/ 10447319509526110 7

    1995

  18. [28]

    A. Lex, N. Gehlenborg, H. Strobelt, R. Vuillemot, and H. Pfister. Upset: Visualization of intersecting sets.IEEE Transactions on Visualization and Computer Graphics, 20(12):1983–1992, 2014. doi: 10.1109/TVCG.2014. 2346248 16

    1983 doi

  19. [29]

    H. Li, Y . Wang, A. Wu, H. Wei, and H. Qu. Structure-aware visualization retrieval. InProceedings of the 2022 CHI Conference on Human Factors in Computing Systems, CHI ’22, article no. 409, 14 pages. Association for Computing Machinery, New York, NY , USA, 2022. doi: 10.1145/3...

    2022 doi

  20. [30]

    W. Li, Z. Wang, Y . Wang, D. Weng, L. Xie, S. Chen, H. Zhang, and H. Qu. Geocamera: Telling stories in geographic visualizations with camera movements. InProceedings of the 2023 CHI Conference on Human Factors in Computing Systems, CHI ’23, article no. 170, 15 pages. Associati...

    2023

  21. [31]

    Y . Ming, H. Qu, and E. Bertini. Rulematrix: Visualizing and understanding classifiers with rules.IEEE Transactions on Visualization and Computer Graphics, 25(1):342–352, 2019. doi: 10.1109/TVCG.2018.2864812 6, 7, 15, 16

    2019

  22. [32]

    Mörth, S

    E. Mörth, S. Bruckner, and N. N. Smit. Scrollyvis: Interactive visual authoring of guided dynamic narratives for scientific scrollytelling.IEEE Transactions on Visualization and Computer Graphics, 29(12):5165–5177,

  23. [33]

    Pandey and A

    S. Pandey and A. Ottley. Benchmarking visual language models on standardized visualization literacy tests.Computer Graphics Forum, 44(3):e70137, 2025. doi: 10.1111/cgf.70137 3

    2025 doi

  24. [34]

    S. Ruan, Q. Guan, P. Griffin, Y . Mao, and Y . Wang. Quantumeyes: To- wards better interpretability of quantum circuits.IEEE Transactions on Visualization and Computer Graphics, 30(9):6321–6333, 2024. doi: 10. 1109/TVCG.2023.3332999 16

    2024

  25. [35]

    Satyanarayan and J

    A. Satyanarayan and J. Heer. Authoring narrative visualizations with ellipsis.Computer Graphics Forum, 33(3):361–370, 2014. doi: 10.1111/ cgf.12392 2

    2014

  26. [36]

    Satyanarayan, D

    A. Satyanarayan, D. Moritz, K. Wongsuphasawat, and J. Heer. Vega-lite: A grammar of interactive graphics.IEEE Transactions on Visualization and Computer Graphics, 23(1):341–350, 2017. doi: 10.1109/TVCG.2016. 2599030 2

    2017 doi

  27. [37]

    Segel and J

    E. Segel and J. Heer. Narrative visualization: Telling stories with data. IEEE Transactions on Visualization and Computer Graphics, 16(6):1139– 1148, 2010. doi: 10.1109/TVCG.2010.179 9

    2010 doi

  28. [38]

    R. Shier. Statistics: 2.2 the wilcoxon signed rank sum test.Mathemat- ics Learning Support Centre. Retrieved from http://www. statstutor. ac. uk/resources/uploaded/wilcoxonsignedranktest. pdf, 7, 2004. 7

    2004

  29. [39]

    M. Shin, A. Tran, S. Wu, A. Mathews, R. Wang, G. Lyall, and L. Xie. Attentionflow: Visualising influence in networks of time series. InPro- ceedings of the 14th ACM International Conference on Web Search and Data Mining, WSDM ’21, 4 pages, p. 1085–1088. Association for Com- pu...

    2021 doi

  30. [40]

    Stoiber, D

    C. Stoiber, D. Ceneda, M. Wagner, V . Schetinger, T. Gschwandtner, M. Streit, S. Miksch, and W. Aigner. Perspectives of visualization on- boarding and guidance in va.Visual Informatics, 6(1):68–83, 2022. doi: 10.1016/j.visinf.2022.02.005 2

    2022 doi

  31. [41]

    Stoiber, F

    C. Stoiber, F. Grassinger, M. Pohl, H. Stitz, M. Streit, and W. Aigner. Visualization onboarding: Learning how to read and use visualizations. In IEEE Workshop on Visualization for Communication. IEEE, 08 2019. doi: 10.31219/osf.io/c38ab 2, 4

    2019 doi

  32. [42]

    Stoiber, M

    C. Stoiber, M. Pohl, and W. Aigner. Design actions for the design of visualization onboarding methods. In2023 IEEE VIS Workshop on Visual- ization Education, Literacy, and Activities (EduVis), pp. 1–10, 2023. doi: 10.1109/EduVis60792.2023.00007 2

    2023

  33. [43]

    Stoiber, S

    C. Stoiber, S. Radkohl, F. Grassinger, D. Moitzi, H. Stitz, E. Goldgruber, D. Girardi, and W. Aigner. Authoring tool for data journalists integrating self-explanatory visualization onboarding concept for a treemap visual- ization. InProceedings of the 15th Biannual Conference ...

    2023

  34. [44]

    Stoiber, C

    C. Stoiber, C. Walchshofer, M. Pohl, B. Potzmann, F. Grassinger, H. Stitz, M. Streit, and W. Aigner. Comparative evaluations of visualization on- boarding methods.Visual Informatics, 6(4):34–50, 2022. doi: 10.1016/j. visinf.2022.07.001 2

    2022 doi

  35. [45]

    Q. Wang, Z. Li, S. Fu, W. Cui, and H. Qu. Narvis: Authoring narrative slideshows for introducing data visualization designs.IEEE Transactions on Visualization and Computer Graphics, PP:1–1, 08 2018. doi: 10.1109/ TVCG.2018.2865232 2

    2018

  36. [46]

    Y . Wang, L. Shen, Z. You, X. Shu, B. Lee, J. Thompson, H. Zhang, and D. Zhang. Wonderflow: Narration-centric design of animated data videos. IEEE Transactions on Visualization and Computer Graphics, 31(9):4638– 4654, 2025. doi: 10.1109/TVCG.2024.3411575 2

    2025

  37. [47]

    Z. J. Wang, R. Turko, and D. H. Chau. Dodrio: Exploring transformer models with interactive visualization. In H. Ji, J. C. Park, and R. Xia, eds., Proceedings of the 59th Annual Meeting of the Association for Computa- tional Linguistics and the 11th International Joint Confere...

    2021 doi

  38. [48]

    X. Wen, T. D. Nguyen, S. Ruan, Q. Shen, J. Sun, F. Zhu, and Y . Wang. Ponzilens+: Visualizing bytecode actions for smart ponzi scheme iden- tification.IEEE Transactions on Visualization and Computer Graphics, 31(9):6451–6465, 2025. doi: 10.1109/TVCG.2024.3516379 6, 15, 16

    2025

  39. [49]

    X. Wen, T. D. Nguyen, L. Zhang, J. Sun, and Y . Wang. Prettismart: Visual interpretation of smart contracts via simulation.IEEE Transactions on Visualization and Computer Graphics, 31(6):3822–3835, 2025. doi: 10. 1109/TVCG.2025.3567130 2, 7, 15, 16

    2025

  40. [50]

    X. Wen, Y . Wang, M. Wu, F. Wang, X. Yue, Q. Shen, Y . Ma, and M. Zhu. <italic>diffseer</italic>: Difference-based dynamic weighted graph visu- alization.IEEE Comput. Graph. Appl., 43(3):12–23, 12 pages, May 2023. doi: 10.1109/MCG.2023.3248289 16

    2023

  41. [51]

    L. Xie, Y . Lin, C. Liu, H. Qu, and X. Shu. Datawink: Reusing and adapting svg-based visualization examples with large multimodal models.IEEE Transactions on Visualization and Computer Graphics, 32(1):824–834,

  42. [52]

    Xu and E

    Z. Xu and E. Wall. Exploring the capability of llms in performing low-level visual analytic tasks on svg data visualizations. In2024 IEEE Visualization and Visual Analytics (VIS), pp. 126–130, 2024. doi: 10.1109/VIS55277. 2024.00033 3

    2024

  43. [53]

    L. Ying, Y . Wang, H. Li, S. Dou, H. Zhang, X. Jiang, H. Qu, and Y . Wu. Reviving static charts into live charts.IEEE Transactions on Visualization and Computer Graphics, 31(8):4314–4328, 2025. doi: 10.1109/TVCG. 2024.3397004 2

    2025

  44. [54]

    X. Zeng, H. Lin, Y . Ye, and W. Zeng. Advancing multimodal large language models in chart question answering with visualization-referenced instruction tuning.IEEE Transactions on Visualization and Computer Graphics, 31(1):525–535, 2025. doi: 10.1109/TVCG.2024.3456159 3

    2025

  45. [55]

    J. Zhao, M. Glueck, F. Chevalier, Y . Wu, and A. Khan. Egocentric analysis of dynamic networks with egolines. InProceedings of the 2016 CHI Conference on Human Factors in Computing Systems, CHI ’16, 12 pages, p. 5003–5014. Association for Computing Machinery, New York, NY , US...

    2016

  46. [56]

    doi: 10.1109/TVCG.2025.3634635 2, 3

    2025

  47. [57]

    J. Zhao, Z. Liu, M. Dontcheva, A. Hertzmann, and A. Wilson. Matrixwave: Visual comparison of event sequence data. InProceedings of the 33rd Annual ACM Conference on Human Factors in Computing Systems, CHI ’15, 10 pages, p. 259–268. Association for Computing Machinery, New York...

    2015

  48. [58]

    J. Zhao, S. Xu, S. Chandrasegaran, C. Bryan, F. Du, A. Mishra, X. Qian, Y . Li, and K.-L. Ma. ChartStory: Automated Partitioning, Layout, and Captioning of Charts into Comic-Style Narratives .IEEE Transactions on Visualization & Computer Graphics, 29(02):1384–1399, Feb. 2023. ...

    2023

  49. [59]

    Y . Zhao, Y . Zhang, Y . Zhang, X. Zhao, J. Wang, Z. Shao, C. Turkay, and S. Chen. Leva: Using large language models to enhance visual analytics. IEEE Transactions on Visualization and Computer Graphics, 31(3):1830– 1847, 2025. doi: 10.1109/TVCG.2024.3368060 2

    2025

  50. [60]

    Q. Zhu, T. Lu, S. Guo, X. Ma, and Y . Yang. Compositingvis: Exploring interactions for creating composite visualizations in immersive environ- ments.IEEE Transactions on Visualization and Computer Graphics, 31(1):591–601, 2025. doi: 10.1109/TVCG.2024.3456210 1

    2025

  51. [61]

    J. Zhao, L. Gou, F. Wang, and M. Zhou. Pearl: An interactive visual analytic tool for understanding personal emotion style derived from social media. In2014 IEEE Conference on Visual Analytics Science and Tech- nology (VAST), pp. 203–212, 2014. doi: 10.1109/V AST.2014.7042496 7, 16

    2014

  52. [62]

    Label mania dashboard

    ZingChart. Label mania dashboard. https://www.zingchart.com/ gallery/label-mania-dashboard . [Online].Accessed: 2026-06-24. 14, 16

    2026

  53. [63]

    Mixed bar and staggered line

    ZingChart. Mixed bar and staggered line. https://www.zingchart. com/gallery/mixed-bar-line-staggered . [Online].Accessed: 2026-06-24. 16

    2026

  54. [64]

    Multiple area charts with shared crosshairs

    ZingChart. Multiple area charts with shared crosshairs. https://www.zingchart.com/gallery/ multiple-area-charts-with-shared-crosshair . [On- line].Accessed: 2026-06-24. 16

    2026

  55. [65]

    Multiple time series scales

    ZingChart. Multiple time series scales. https://www.zingchart.com/ gallery/multiple-time-series-scales . [Online].Accessed: 2026- 06-24. 16 A APPENDICES Section A.1 details the Semantic Inference prompt, which reconstructs structured explanation units from multimodal inputs. S...

    2026

  56. [66]

    Zhu-Tian, Y

    C. Zhu-Tian, Y . Wang, Q. Wang, Y . Wang, and H. Qu. Towards automated infographic design: Deep learning-based auto-extraction of extensible timeline.IEEE Transactions on Visualization and Computer Graphics, 26(1):917–926, 2020. doi: 10.1109/TVCG.2019.2934810 2

    2020

  57. [71]

    Structural- Describe the overall composition and layout pattern (e.g., small multiples, coordinated views, overlays); explain shared elements (shared axes, legends) and how they connect different visual components

  58. [72]

    Visual Encoding- Explain how data attributes are mapped to visual channels (position, color, size, shape, or mark type); focus on how to read each individual mark within a visual component

  59. [73]

    Data- Identify what data is shown in each visual compo- nent: variables, units, categories, and time ranges

  60. [74]

    Analytical- Describe observable patterns, trends, compar- isons, or relationships

  61. [75]

    Look at the colored dots

    Interaction- Infer supported interactions (hover, filter, brush, zoom) and explain what analytical tasks they en- able. Constraints: • Knowledgemustbe grounded in visible evidence • Analytical knowledgemustdescribe observable patterns • Interaction knowledgemustonly be include...

  62. [2012]

    doi: 10.1109/PacificVis.2012.6183556 1, 7

    2012

  63. [2023]

    doi: 10.1109/TVCG.2022.3205769 2

    2022

  64. [2025]

    doi: 10.1111/cgf.70107 2

    doi

  65. [2026]

    doi: 10.1016/j.xinn.2025.101253 7

    2025

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.