Pith. sign in

REVIEW 4 major objections 6 minor 71 references

A Multimodal Framework for Understanding Collaborative Design Processes

T0 review · 4 major / 6 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read The paper claims that collaborative design workshops can be recorded across video, audio, notes, and gaze and distilled into editable topic cards that let analysts reconstruct why design decisions were made.

desk verdict Practical, openly available multimodal workshop-analysis system; the unvalidated segmentation thresholds in Sec 4.3 are the main soft spot—still deserves peer review. read the letter →

arxiv 2508.06117 v1 pith:UH2H7Y7F submitted 2025-08-08 cs.HC

Maurice Koch , Nelusa Pathmanathan , Daniel Weiskopf , Kuno Kurzhals This is my paper
classification cs.HC
keywords collaborativedesignworkshopsmultimodaldataanalysisvisualanalyticstopicsegmentationeyetrackingstudymethodologyworkshopAI-assistedartifactextraction
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

Collaborative design workshops usually leave only consolidated notes and final artifacts, so the reasoning that produced a decision is lost. This paper tries to close that gap with a modular framework that records workshops across video, audio, notes, and gaze, then automatically extracts artifacts such as activity, attention, transcripts, and topic segments. The centerpiece is reCAPit, an interactive visual analysis system whose two-step segmentation pipeline turns hours of raw multimodal data into editable topic cards that analysts can filter, annotate, and assemble into a traceable summary. If the approach works as claimed, workshop organizers can study the process that produced an outcome rather than only the outcome, and can communicate that process transparently.

What carries the argument

The load-bearing mechanism is a two-step multimodal segmentation that converts continuous recordings into discrete, editable topic segments. Step one applies pruned exact linear time (PELT) change-point detection to the multivariate time series of activity or visual attention across predefined areas of interest; step two refines each coarse segment by splitting the audio transcript on 1.5-second pauses and introducing a new boundary when the cosine similarity between consecutive text-embedding chunks falls below 0.5. These segments are surfaced as topic cards, the unit that carries all downstream analysis—timeline overview, keyword filtering, and authoring—so every higher-level conclusion in

What would settle it

Take one recorded workshop from the paper's corpus, have independent coders mark where they believe the discussion topic changes, and compare those boundaries with the pipeline's output; then rerun the segmentation with the penalty, gap, and cosine thresholds varied over a few plausible values and count how many topic-card contents change. Agreement at chance level, or large reshuffling under small threshold changes, would falsify the framework's entry-point claim.

Watch

Extended reading notes

Core claim

The paper claims that collaborative design workshops can be recorded and made analyzable through a two-step multimodal segmentation pipeline plus interactive visualization. Activity or attention over predefined areas is segmented with the PELT algorithm; each interval is refined by splitting transcripts at pauses and adding a boundary when cosine similarity between consecutive text-embedding chunks drops below a threshold, marking a topic shift. The resulting topic cards carry titles, quotes, screenshots, notes, and per-segment statistics, allowing analysts to reconstruct how a requirement or concept entered the discussion, including statements whose meaning depends on what someone was point

Load-bearing premise

The entire analysis inherits the assumption that the hand-set thresholds and the activity/attention detectors produce segment boundaries that correspond to the topics participants actually discussed, a premise the paper does not validate against human-coded ground truth or sensitivity tests.

Editorial extensions

If this is right

  • Workshop organizers can move from final artifacts to process: the two case studies trace specific requirements and concepts back to who introduced them, when, and with what evidence.
  • The roughly 1–2 hour AI-driven processing window opens recorded data to analysts who previously skipped video or gaze because coding-based analysis took days.
  • Topic cards, with quotes, screenshots, notes, and statistics, provide a reusable summary format for transparently disseminating how results were derived.
  • The modular design means workshops can record only a subset of modalities, such as video and audio alone, and still obtain the same card-based analysis structure.

Reading between the lines

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

  • The segmentation scheme is a plausible template for other co-located collaborative settings, such as classroom group work, design sprints, or emergency-team training, but each setting would require validating the thresholds anew.
  • A natural next step would be adaptive thresholds: an analyst's edits to a topic card could serve as feedback to recalibrate the PELT penalty or cosine threshold for subsequent workshops.
  • The paper's indexicality example suggests a concrete testable extension: use gaze and video to automatically resolve deictic expressions like 'this' in 'this ensures the city becomes fairer', replacing the manual linkage the analyst currently performs.
  • If topic cards are treated as authored annotations rather than machine outputs, they could accumulate into a corpus for supervised learning of topic boundaries, closing the loop between human analysis and automatic extraction.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 6 minor

Summary. The paper proposes a modular framework for capturing, analyzing, and communicating collaborative design workshops. It describes a data-acquisition setup (video, audio, gaze, notes), automatic extraction of artifacts (activity, visual attention, transcripts, topic segments), and an interactive visual analysis system called reCAPit. The central technical contribution is a two-step segmentation pipeline: initial segmentation via PELT change-point detection on activity/attention time series, followed by refinement using text-embedding cosine similarity to detect topic shifts. The system is evaluated through two case studies—a band-practice visualization workshop and an urban-planning workshop—with expert interviews, and the paper provides methodological recommendations based on six workshops. The authors claim that the framework enables analysts to move beyond final artifacts and trace the collaborative process that led to design decisions.

Significance. If the central claim holds, the work addresses a real gap: workshop analysis today relies heavily on manual coding and final artifacts, and the proposed framework offers a concrete pathway toward transparent, multimodal process analysis. The strengths are substantial: the framework is modular and flexible; the system is publicly available with source code; the paper reports on six workshops with two detailed case studies; and it engages seriously with rigor criteria from the design-study literature (Meyer and Dykes). The expert interviews provide ecologically valid feedback and show genuine analytical value, e.g., in the urban-planning case the expert found the combination of video, heatmap, and transcript useful for resolving indexicality. However, the load-bearing automatic segmentation pipeline is presented without quantitative validation, which risks undermining the reliability of every downstream artifact. The paper's own admissions—the expert's request to adjust topic-segment boundaries (§6.2) and the concession that events can be missed (§8)—indicate that this is a real rather than hypothetical concern.

major comments (4)
  1. [§4.3, Topic Segmentation] The thresholds β=10 (PELT penalty), gap=1.5 s, and cosine similarity=0.5 are described only as 'empirically found' with no accuracy assessment, sensitivity analysis, or comparison to human-annotated topic boundaries. The output of this pipeline feeds every downstream artifact (topic cards, keyword filtering, timeline shading, case-study narratives), so a misplaced boundary can merge or split topics and silently mislead analysis. The paper's own evidence supports this concern: §6.2 reports that the urban-planning expert requested the ability to adjust topic-segment start/end timestamps, and §8 concedes that events can be missed. Please add a validation of the segmentation (e.g., agreement with human annotations on a subset, or sensitivity to each threshold) or, if no validation is feasible, make the caveat prominent and treat boundary correction as an essential—not optional—workflow in th
  2. [§4.5, Transcript and Roles] The paper states that speaker diarization was performed manually. This is a substantial human step in the 'AI-based artifact extraction' pipeline, and it is load-bearing for role-based timelines, summary statistics in topic cards, and quote attribution in the case studies (§6.1). The paper should report the time and effort required for manual diarization in the two workshops and discuss how the framework scales to longer or larger-group workshops. If the manual step is essential, the characterization of the pipeline as 'AI-based artifact extraction' should be qualified accordingly.
  3. [§4.2, Activity] The activity signal combines background subtraction and hand tracking, but no validation is reported. The multimodal streamgraph and topic-card statistics rely on this signal; the case study in §6.1 interprets activity in Randy's area as pointing to a calendar sketch, but it is not established that the detector's output corresponds reliably to such gestures. At minimum, provide a basic evaluation (e.g., frame-level comparison against manual annotation for a short segment) or explicitly delimit what the activity measure can and cannot support. Without this, the visual analysis and the case-study interpretations built on activity peaks are vulnerable to detector artifacts.
  4. [§7.1 and Case Studies] The claim that 'our mainly AI-driven data processing can be achieved within 1–2 hours' is stated without supporting measurements or comparison to the alternative of open coding. The two case studies demonstrate usage but do not quantify recall of important events, time savings, or agreement with the expert's established analysis workflow. To make the efficiency claim credible, please provide evidence for the time estimate and, if possible, a structured comparison of the system's outputs against the urban-planning expert's prior coding results.
minor comments (6)
  1. [Throughout] The text contains several corrupted glyphs (e.g., '1�5 seconds', '0�5', and entries in Table 1) that obscure thresholds and data-source descriptions. These should be corrected in the camera-ready version.
  2. [§4.3] The interval notation '(ci, ci+1)' is rendered incorrectly; please clarify how the initial segments are defined and how the refined change point c̃ is placed within them.
  3. [§5] The source code URL is missing/placeholder in the provided version; include a working link in the final manuscript. Also ensure the supplemental material is cited consistently with the data repository.
  4. [References] Reference [14] (Dynamicland) has an inaccessible URL; provide a full, verifiable citation.
  5. [Figures] Figure 7 is described as a 15-minute recording, but the streamgraph and topic cards do not show a time axis; consider adding time labels to orient the reader. In Figure 8, the requirement list is difficult to read in grayscale; enhance contrast or enlarge the relevant region.
  6. [§6.1] The inline notation 'Segmentation [(temporal) of a recording session is necessary]' is awkward; rephrase to clarify that this is the requirement text from the workshop outcome.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: artifact extraction and case-study analysis are self-contained and not reduced to the system's own outputs

full rationale

The paper's derivation chain is a pipeline: recorded multimodal data is processed by external algorithms (PELT [26], Whisper [41], GTE [59], MediaPipe [58], background subtraction [61]) into segments, transcripts, and statistics, and these artifacts are then visualized in reCAPit. No claimed result is defined in terms of the system's own output. The Section 4.3 thresholds (β=10, gap=1.5 s, cosine=0.5) are free parameters chosen 'empirically' rather than fitted to a target quantity that the paper later claims to predict; the case studies are qualitative demonstrations using expert interviews (Pair Analytics, semi-structured interview), not holdout predictions forced by those parameters. The author self-citations used in the design rationale ([29] supplemental material, [31], [33], [52], [53]) are contextual and not load-bearing: the central segmentation and summarization claims rest on independently published methods and on the authors' own reported interviews. The paper's acknowledged limitations—the urban-planning expert's request for adjustable topic-segment boundaries (§6.2) and the §8 admission that 'there is a likelihood of missing events'—are soundness concerns about threshold choice and coverage, not evidence that the pipeline's outputs were fed back as inputs. Accordingly, no circular step meeting the required quote-and-reduction standard is present.

Assumptions & free parameters 3 free parameters · 6 assumptions · 0 invented entities

The paper's contributions rest on a pipeline whose key thresholds and detection quality are assumed, not validated. The three free parameters in Section 4.3 directly control segmentation; the detection axioms (activity, attention, transcription) are inherited from off-the-shelf tools without domain-specific evaluation. No new physical or conceptual entities are postulated.

free parameters (3)
  • PELT penalty constant beta = 10
    Empirically chosen in Section 4.3 for 'sufficiently coarse segmentation'; no sensitivity analysis.
  • Gap threshold for transcript chunks = 1.5 s
    Chosen in Section 4.3 to account for fast-paced group discussions; no quantitative justification.
  • Cosine similarity threshold for topic shifts = 0.5
    Chosen in Section 4.3; no evaluation of threshold effects on segmentation quality.
assumptions (6)
  • domain assumption Detected change points in activity/attention time series (via PELT) correspond to meaningful workshop phases.
    Section 4.3 builds the whole segmentation pipeline on PELT without validating that activity/attention change points align with topic boundaries.
  • domain assumption Semantic similarity of transcript chunks (GTE embeddings) detects topic shifts.
    Section 4.3 uses cosine similarity below 0.5 as a topic boundary proxy; no comparison to human-annotated topic segments.
  • domain assumption Fixation hits on predefined AOIs measure visual attention, and joint attention predicts collaboration quality.
    Section 4.4 adopts standard eye-tracking methodology relying on reference [44] for the link to collaboration.
  • domain assumption Whisper transcription and manual speaker assignment are sufficiently accurate for analysis.
    Section 4.5 acknowledges manual diarization; no error analysis for transcription.
  • domain assumption Background subtraction plus hand tracking activity reflects meaningful activity in AOIs.
    Section 4.2 states it 'can still facilitate detecting phases of individual and joint interactions', without evaluation.
  • domain assumption LLM-generated topic titles accurately summarize segment content.
    Section 5.4 uses GPT-4 titles; no accuracy assessment.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Multimodal Framework for Understanding Collaborative Design Processes." pith.science (2026). https://pith.science/paper/UH2H7Y7F

@misc{pith2026250806117,
  author       = {Pith},
  title        = {Pith review of: A Multimodal Framework for Understanding Collaborative Design Processes},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UH2H7Y7F}},
  note         = {Machine review of arXiv:2508.06117}
}
read the original abstract

An essential task in analyzing collaborative design processes, such as those that are part of workshops in design studies, is identifying design outcomes and understanding how the collaboration between participants formed the results and led to decision-making. However, findings are typically restricted to a consolidated textual form based on notes from interviews or observations. A challenge arises from integrating different sources of observations, leading to large amounts and heterogeneity of collected data. To address this challenge we propose a practical, modular, and adaptable framework of workshop setup, multimodal data acquisition, AI-based artifact extraction, and visual analysis. Our interactive visual analysis system, reCAPit, allows the flexible combination of different modalities, including video, audio, notes, or gaze, to analyze and communicate important workshop findings. A multimodal streamgraph displays activity and attention in the working area, temporally aligned topic cards summarize participants' discussions, and drill-down techniques allow inspecting raw data of included sources. As part of our research, we conducted six workshops across different themes ranging from social science research on urban planning to a design study on band-practice visualization. The latter two are examined in detail and described as case studies. Further, we present considerations for planning workshops and challenges that we derive from our own experience and the interviews we conducted with workshop experts. Our research extends existing methodology of collaborative design workshops by promoting data-rich acquisition of multimodal observations, combined AI-based extraction and interactive visual analysis, and transparent dissemination of results.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

71 extracted references · 35 canonical work pages

  1. [1]

    Arias-Hernandez, L

    R. Arias-Hernandez, L. T. Kaastra, T. M. Green, and B. Fisher. Pair Ana- lytics: Capturing Reasoning Processes in Collaborative Visual Analytics. In Proceedings of the IEEE Hawaii International Conference on System Sciences, 2011. doi: 10.1109/HICSS.2011.339 7, 9

    work page doi:10.1109/hicss.2011.339 2011

  2. [2]

    Banerjee, M

    M. Banerjee, M. Capozzoli, L. McSweeney, and D. Sinha. Beyond Kappa: A Review of Interrater Agreement Measures. Canadian Journal of Statis- tics, 27(1):3–23, 1999. doi: 10.2307/3315487 4

    work page doi:10.2307/3315487 1999

  3. [3]

    Batch, Y

    A. Batch, Y . Ji, M. Fan, J. Zhao, and N. Elmqvist. uxSense: Supporting User Experience Analysis with Visualization and Computer Vision. IEEE Transactions on Visualization and Computer Graphics, 30(7):3841–3856,

    work page

  4. [4]

    Bergstrom and K

    T. Bergstrom and K. Karahalios. Conversation Clock: Visualizing au- dio patterns in co-located groups. In Proceedings of the IEEE Hawaii International Conference on System Sciences, pp. 78–78, 2007. doi: 10. 1109/HICSS.2007.151 2

    work page 2007

  5. [5]

    Berkovsky, R

    S. Berkovsky, R. Taib, I. Koprinska, E. Wang, Y . Zeng, J. Li, and S. Kleit- man. Detecting Personality Traits Using Eye-tracking Data. In Proceed- ings of the ACM Conference on Human Factors in Computing Systems, pp. 221:1–221:12, 2019. doi: 10.1145/3290605.3300451 9

    arXiv 2019

  6. [6]

    Blascheck, M

    T. Blascheck, M. John, K. Kurzhals, S. Koch, and T. Ertl. V A2: A Visual Analytics Approach for Evaluating Visual Analytics Applications. IEEE Transactions on Visualization and Computer Graphics, 22(1):61–70, 2016. doi: 10.1109/TVCG.2015.2467871 2, 9

    arXiv 2016

  7. [7]

    Blikstein

    P. Blikstein. Multimodal Learning Analytics. In Proceedings of the ACM Conference on International Learning Analytics & Knowledge, p. 102–106,

    work page

  8. [8]

    J. Boy, F. Detienne, and J.-D. Fekete. Storytelling in Information Visualiza- tions: Does It Engage Users to Explore Data? In Proceedings of the ACM Conference on Human Factors in Computing Systems , pp. 1449–1458,

    work page

Show all 71 references
  1. [9]

    Burch, A

    M. Burch, A. Kull, and D. Weiskopf. AOI Rivers for Visualizing Dynamic Eye Gaze Frequencies. Computer Graphics Forum, 32(3):281–290, 2013. doi: 10.1111/cgf.12115 2, 5

    2013 doi

  2. [10]

    Byron and M

    L. Byron and M. Wattenberg. Stacked Graphs – Geometry & Aesthetics. IEEE Transactions on Visualization and Computer Graphics, 14(6):1245– 1252, 2008. doi: 10.1109/tvcg.2008.166 2

    2008 doi

  3. [11]

    Cavallo, M

    M. Cavallo, M. Dolakia, M. Havlena, K. Ocheltree, and M. Podlaseck. Immersive Insights: A Hybrid Analytics System for Collaborative Ex- ploratory Data Analysis. In Proceedings of the ACM Symposium on Virtual Reality Software and Technology , pp. 9:1–9:12, 2019. doi: 10. 1145/3...

    2019

  4. [12]

    J. W. Cresswell. Qualitative Inquiry and Research Design: Choosing Among Five Traditions. Sage Publications, 1998. 1, 2, 9

    1998

  5. [13]

    Dasu, Y .-H

    K. Dasu, Y .-H. Kuo, and K.-L. Ma. Character-Oriented Design for Visual Data Storytelling. IEEE Transactions on Visualization and Computer Graphics, 30(1):98–108, 2024. doi: 10.1109/tvcg.2023.3326578 3

    2024

  6. [14]

    Dynamicland, 2025

    Dynamicland. Dynamicland, 2025. Available: �������������������� ���. Accessed: 2025-08-07. 9

    2025

  7. [15]

    Echeverria, R

    V . Echeverria, R. Martinez-Maldonado, and S. Buckingham Shum. To- wards Collaboration Translucence: Giving Meaning to Multimodal Group Data. In Proceedings of the ACM Conference on Human Factors in Com- puting Systems, pp. 39:1–39:16, 2019. doi: 10.1145/3290605.3300269 2

    2019

  8. [16]

    El-Assady, V

    M. El-Assady, V . Gold, C. Acevedo, C. Collins, and D. Keim. ConToVi: Multi-Party Conversation Exploration using Topic-Space Views.Computer Graphics Forum, 35(3):431–440, 2016. doi: 10.1111/cgf.12919 2

    2016 doi

  9. [17]

    B. Ens, S. Goodwin, A. Prouzeau, F. Anderson, F. Y . Wang, S. Gratzl, Z. Lucarelli, B. Moyle, J. Smiley, and T. Dwyer. Uplift: A Tangible and Immersive Tabletop System for Casual Collaborative Visual Analytics. IEEE Transactions on Visualization and Computer Graphics, 27(2):11...

    2021

  10. [18]

    Fouse, N

    A. Fouse, N. Weibel, E. Hutchins, and J. D. Hollan. ChronoViz: A System for Supporting Navigation of Time-coded Data. In Extended Abstracts of the ACM Conference on Human Factors in Computing Systems , pp. 299–304, 2011. doi: 10.1145/1979742.1979706 2

    2011

  11. [20]

    Fulda, M

    J. Fulda, M. Brehmer, and T. Munzner. TimeLineCurator: Interactive Authoring of Visual Timelines from Unstructured Text.IEEE Transactions on Visualization and Computer Graphics, 22(1):300–309, 2016. doi: 10. 1109/TVCG.2015.2467531 3

    2016

  12. [21]

    Havre, B

    S. Havre, B. Hetzler, and L. Nowell. ThemeRiver: Visualizing Theme Changes over Time. In IEEE Symposium on Information Visualization, pp. 115–123, 2000. doi: 10.1109/INFVIS.2000.885098 2

  13. [22]

    Huang, Q

    Z. Huang, Q. He, K. T. Maher, X. Deng, Y . Lai, C. Ma, S. F. Qin, Y . Liu, and H. Wang. SpeechMirror: A Multimodal Visual Analytics System for Personalized Reflection of Online Public Speaking Effectiveness. IEEE Transactions on Visualization and Computer Graphics, 30(1):606–616,

  14. [23]

    Isenberg, T

    P. Isenberg, T. Zuk, C. Collins, and S. Carpendale. Grounded Evaluation of Information Visualizations. In Proceedings of the ACM Workshop on Evaluation and Beyond – Methodological Approaches to Visualization, pp. 6:1–6:8, 2008. doi: 10.1145/1377966.1377974 2

    2008

  15. [24]

    Social Gaze Space

    M. Jording, A. Hartz, G. Bente, M. Schulte-Rüther, and K. V ogeley. The “Social Gaze Space”: A Taxonomy for Gaze-Based Communication in Triadic Interactions. Frontiers in Psychology, 9:226:1–226:8, 2018. doi: 10.3389/fpsyg.2018.00226 2

    2018

  16. [25]

    Kerzner, S

    E. Kerzner, S. Goodwin, J. Dykes, S. Jones, and M. Meyer. A Framework for Creative Visualization-Opportunities Workshops. IEEE Transactions on Visualization and Computer Graphics, 25(1):748–758, 2019. doi: 10. 1109/tvcg.2018.2865241 1, 2

    2019

  17. [26]

    doi: 10.1109/TVCG.2023.3326932 2

    2023

  18. [27]

    Y . Kim, S. Butail, M. Tscholl, L. Liu, and Y . Wang. An Exploratory Ap- proach to Measuring Collaborative Engagement in Child Robot Interaction. In Proceedings of the ACM Conference on International Learning Ana- lytics & Knowledge, pp. 209–217, 2020. doi: 10.1145/3375462.3375522 5

    2020

  19. [28]

    Knoll, A

    C. Knoll, A. Cetin, T. Möller, and M. Meyer. Extending Recommendations for Creative Visualization-Opportunities Workshops. InProceedings of the IEEE Workshop on Evaluation and Beyond – Methodological Approaches to Visualization, pp. 81–88, 2020. doi: 10.1109/beliv51497.2020.00017 3

    2020

  20. [29]

    A Multimodal Framework for Understanding Collaborative Design Processes

    M. Koch, N. Pathmanathan, D. Weiskopf, and K. Kurzhals. Supplemental Material for “A Multimodal Framework for Understanding Collaborative Design Processes”, 2025. DaRUS: The Data Repository of the University of Stuttgart. doi: 10.18419/DARUS-5166 5, 7

    2025 doi

  21. [30]

    Killick, P

    R. Killick, P. Fearnhead, and I. A. Eckley. Optimal Detection of Change- points With a Linear Computational Cost. Journal of the American Sta- tistical Association, 107(500):1590–1598, 2012. doi: 10.1080/01621459. 2012.737745 4

    2012

  22. [31]

    Kurzhals

    K. Kurzhals. Anonymizing Eye-tracking Stimuli with Stable Diffusion. Computers & Graphics, 119(1):103898:1–103898:9, 2024. doi: 10.1016/j. cag.2024.103898 9

    2024

  23. [32]

    Kurzhals, M

    K. Kurzhals, M. Hlawatsch, F. Heimerl, M. Burch, T. Ertl, and D. Weiskopf. Gaze Stripes: Image-Based Visualization of Eye Tracking Data. IEEE Transactions on Visualization and Computer Graphics, 22(1):1005–1014,

  24. [33]

    Kurzhals, N

    K. Kurzhals, N. Rodrigues, M. Koch, M. Stoll, A. Bruhn, A. Bulling, and D. Weiskopf. Visual Analytics and Annotation of Pervasive Eye Tracking Video. In Proceedings of the ACM Symposium on Eye Tracking Research & Applications, pp. 16:1–16:9, 2020. doi: 10.1145/3379155.3391326 2

    2020

  25. [34]

    Kong and Y

    Y . Kong and Y . Fu. Human Action Recognition and Prediction: A Survey. International Journal of Computer Vision, 130(5):1366–1401, 2022. doi: 10.1007/s11263-022-01594-9 9

    2022 doi

  26. [35]

    Liebers, S

    C. Liebers, S. Agarwal, M. Krug, K. Pitsch, and F. Beck. VisCoMET: Visually Analyzing Team Collaboration in Medical Emergency Trainings. Computer Graphics Forum, 42(3):149–160, 2023. doi: 10.1111/cgf.14819 2

    2023 doi

  27. [36]

    NVivo, 2025

    Lumivero (formerly QSR International). NVivo, 2025. Available: ������ ����������������������������� . Accessed: 2025-08-07. 2 10 © 2025 IEEE. This is the author’s version of the article that has been published in IEEE Transactions on Visualization and Computer Graphics. The fi...

    2025

  28. [37]

    Mahyar and M

    N. Mahyar and M. Tory. Supporting Communication and Coordination in Collaborative Sensemaking. IEEE Transactions on Visualization and Computer Graphics, 20(12):1633–1642, 2014. doi: 10.1109/TVCG.2014. 2346573 9

    2014 doi

  29. [38]

    Martinez-Maldonado, V

    R. Martinez-Maldonado, V . Echeverria, G. Fernandez Nieto, and S. Buck- ingham Shum. From Data to Insights: A Layered Storytelling Approach for Multimodal Learning Analytics. In Proceedings of the ACM Confer- ence on Human Factors in Computing Systems, pp. 21:1–21:15, 2020. do...

    2020

  30. [39]

    H. Lam, E. Bertini, P. Isenberg, C. Plaisant, and S. Carpendale. Empirical Studies in Information Visualization: Seven Scenarios. IEEE Transactions on Visualization and Computer Graphics, 18(9):1520–1536, 2012. doi: 10 .1109/TVCG.2011.279 9

    2012

  31. [40]

    Pirolli and S

    P. Pirolli and S. Card. The Sensemaking Process and Leverage Points for Analyst Technology as Identified Through Cognitive Task Analysis. In Proceedings of the IEEE International Conference on Intelligence Analysis, pp. 2–4, 2005. 9

    2005

  32. [41]

    Radford, J

    A. Radford, J. W. Kim, T. Xu, G. Brockman, C. Mcleavey, and I. Sutskever. Robust Speech Recognition via Large-Scale Weak Supervision. In Pro- ceedings of the International Conference on Machine Learning, pp. 28492– 28518, 2023. 4

    2023

  33. [42]

    D. C. Richardson and R. Dale. Looking to Understand: The Coupling between Speakers’ and Listeners’ Eye Movements and Its Relationship to Discourse Comprehension. Cognitive Science, 29(6):1045–1060, 2005. doi: 10.1207/s15516709cog0000_29 5

    2005 doi

  34. [43]

    Rogers, A

    J. Rogers, A. H. Patton, L. Harmon, A. Lex, and M. Meyer. Insights From Experiments With Rigor in an EvoBio Design Study. IEEE Transactions on Visualization and Computer Graphics, 27(2):1106–1116, 2021. doi: 10 .1109/tvcg.2020.3030405 2

    2021

  35. [44]

    Meyer and J

    M. Meyer and J. Dykes. Criteria for Rigor in Visualization Design Study. IEEE Transactions on Visualization and Computer Graphics, 26(1):87–97,

  36. [45]

    Sedlmair, M

    M. Sedlmair, M. Meyer, and T. Munzner. Design Study Methodology: Reflections from the Trenches and the Stacks. IEEE Transactions on Visualization and Computer Graphics, 18(12):2431–2440, 2012. doi: 10. 1109/TVCG.2012.213 1, 2, 7

    2012

  37. [46]

    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 3, 6

    2010 doi

  38. [47]

    Y . Shi, X. Lan, J. Li, Z. Li, and N. Cao. Communicating with Motion: A Design Space for Animated Visual Narratives in Data Videos. In Proceedings of the ACM Conference on Human Factors in Computing Systems, pp. 605:1–605:13, 2021. doi: 10.1145/3411764.3445337 2

    2021

  39. [48]

    E. J. Soure, E. Kuang, M. Fan, and J. Zhao. CoUX: Collaborative Visual Analysis of Think-Aloud Usability Test Videos for Digital Interfaces.IEEE Transactions on Visualization and Computer Graphics, 28(1):643–653,

  40. [49]

    T. Tang, Y . Wu, Y . Wu, L. Yu, and Y . Li. VideoModerator: A Risk-aware Framework for Multimodal Video Moderation in E-Commerce. IEEE Transactions on Visualization and Computer Graphics, 28(1):846–856,

  41. [50]

    Schneider, K

    B. Schneider, K. Sharma, S. Cuendet, G. Zufferey, P. Dillenbourg, and R. Pea. Leveraging Mobile Eye-trackers to Capture Joint Visual Attention in Co-located Collaborative Learning Groups. International Journal of Computer-Supported Collaborative Learning, 13(3):241–261, 2018. ...

    2018 doi

  42. [51]

    MAXQDA, 2025

    VERBI GmbH. MAXQDA, 2025. Available: ���������������������� . Accessed: 2025-08-07. 2

    2025

  43. [52]

    Weiskopf

    D. Weiskopf. Vis4Vis: Visualization for (Empirical) Visualization Re- search. In Foundations of Data Visualization , pp. 209–224. Springer International Publishing, 2020. doi: 10.1007/978-3-030-34444-3_10 3

    2020 doi

  44. [53]

    Weiskopf

    D. Weiskopf. Bridging Quantitative and Qualitative Methods for Visu- alization Research: A Data/Semantics Perspective in Light of Advanced AI. In Proceedings of the IEEE Workshop on Evaluation and Beyond – Methodological Approaches for Visualization, pp. 119–128, 2024. doi: 10...

    2024

  45. [54]

    Wittenburg, H

    P. Wittenburg, H. Brugman, A. Russel, A. Klassmann, and H. Sloetjes. ELAN: A Professional Framework for Multimodality Research. In Pro- ceedings of the ELRA International Conference on Language Resources and Evaluation, pp. 1556–1559, 2006. Max Planck Institute for Psy- cholin...

    2006

  46. [55]

    K. K. Wong, X. Wang, Y . Wang, J. He, R. Zhang, and H. Qu. Anchorage: Visual Analysis of Satisfaction in Customer Service Videos Via Anchor Events. IEEE Transactions on Visualization and Computer Graphics , 30(7):4008–4022, 2024. doi: 10.1109/TVCG.2023.3245609 2

    2024

  47. [56]

    Wu and H

    A. Wu and H. Qu. Multimodal Analysis of Video Collections: Visual Exploration of Presentation Techniques in TED Talks. IEEE Transactions on Visualization and Computer Graphics, 26(7):2429–2442, 2020. doi: 10 .1109/TVCG.2018.2889081 2

    2020

  48. [57]

    doi: 10.1109/TVCG.2021.3114781 2

    2021

  49. [58]

    Tonsen, C

    M. Tonsen, C. K. Baumann, and K. Dierkes. A High-Level Description and Performance Evaluation of Pupil Invisible. arXiv:2009.00508 [cs.CV],

    2009 arXiv

  50. [59]

    doi: 10.48550/arXiv.2009.00508 3

    2009 doi arXiv

  51. [60]

    Zhao and N

    Z. Zhao and N. Elmqvist. The Stories We Tell about Data: Surveying Data-driven Storytelling Using Visualization. IEEE Computer Graphics and Applications, 43(4):97–110, 2023. doi: 10.1109/mcg.2023.3269850 3

    2023

  52. [61]

    Zivkovic and F

    Z. Zivkovic and F. van der Heijden. Efficient Adaptive Density Estimation per Image Pixel for the Task of Background Subtraction. Pattern Recog- nition Letters, 27(7):773–780, 2006. doi: 10.1016/j.patrec.2005.11.005 4 11

    2006 doi

  53. [66]

    H. Zeng, X. Wang, A. Wu, Y . Wang, Q. Li, A. Endert, and H. Qu. EmoCo: Visual Analysis of Emotion Coherence in Presentation Videos. IEEE Transactions on Visualization and Computer Graphics, 26(1):927–937,

  54. [67]

    doi: 10.1109/TVCG.2019.2934656 2

    2019

  55. [68]

    Zhang, V

    F. Zhang, V . Bazarevsky, A. Vakunov, A. Tkachenka, G. Sung, C.-L. Chang, and M. Grundmann. MediaPipe Hands: On-device Real-time Hand Tracking. arXiv:2006.10214 [cs.CV], 2020. doi: 10.48550/arXiv. 2006.10214 4

    2006 doi arXiv

  56. [69]

    Zhang, Y

    X. Zhang, Y . Zhang, D. Long, W. Xie, Z. Dai, J. Tang, H. Lin, B. Yang, P. Xie, F. Huang, M. Zhang, W. Li, and M. Zhang. mGTE: Generalized Long-Context Text Representation and Reranking Models for Multilingual Text Retrieval. In Proceedings of the ACL Conference on Empirical M...

  57. [70]

    doi: 10.18653/v1/2024.emnlp-industry.103 4

    2024 doi

  58. [2013]

    doi: 10.1145/2460296.2460316 2

  59. [2015]

    doi: 10.1145/2702123.2702452 3

  60. [2016]

    doi: 10.1109/tvcg.2015.2468091 3

    2015

  61. [2020]

    doi: 10.1109/TVCG.2019.2934539 2

    2019

  62. [2022]

    doi: 10.1109/TVCG.2021.3114822 2

    2021

  63. [2024]

    doi: 10.1109/TVCG.2023.3241581 2, 6

    2023

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.