Shuichiro Ogawa
日本語

Notes

Research, experiments, and reflections on AI and design, published along with the process of making them.

Back to the monthly list

Visualization and art (13)

Data visualization, data art, and generative art.

  • A literature map for a review that reads data visualization as an operation that changes the angle of interpretation

    2026-10-07

    To write a review paper on the hypothesis that "data visualization is one of the operations for exploring the interpretation of data from new angles," we collected literature in four domains: exploratory data analysis and visual analytics, the cognitive science of external representations, design theory (framing, ambiguity, fixation), and the humanities and critical visualization (118 references, 67 with full text obtained). Following the novelty conditions of paper-pipeline (Route B of the RW axis), we recorded the nearest reviews in each domain and their shelves, a mapping table onto five new shelves set along the axis of the hypothesis (taking over computation, discovery within a frame, frame restructuring, frame locking, and recording and auditing paths), five candidate splits and joins that cross two or more domains, the prospects for the removal and substitution tests, a candidate one-sentence statement of the newly enabled judgment, the nearest work in three directions, literature that runs against the hypothesis, and the status of the eight conditions together with the next tasks. Because the claim that "changing a representation changes the way people think" can already be stated in the vocabulary of Scaife and Rogers, Kirsh, and Tversky, the novelty must be placed not in a claim about the phenomenon but in the restructuring of the literature's shelves. A note prepared for a paper.

  • What Does Data Visualization Show When It Departs from Efficiency? Wholes Without Numbers, Readings That Stay Open, and Deliberate Abstraction

    2026-10-07

    Moving away from the earlier notes that measured the value of data visualization by optimization criteria such as speed, accuracy, behavior change, and number of users, this note examines the significance of seeing the whole without numbers, of opening new angles because the way of reading is not fixed, and of abstracting deliberately to avoid locally optimal views. It draws on academic research (visualization research, design theory in HCI, the sociology of data, and the anthropology of audit) and on statements by people who turn data into images. The mainstream yardstick of value in visualization research goes back to an economic model that measures value by effectiveness and efficiency (van Wijk 2005), and a recent critique argues that the canon embeds universality, objectivity, and efficiency as values (Saharan et al. 2026, not peer reviewed). A heatmap that ranks lowest for the precision of reading individual values can be the most useful big-picture view (Bertini et al. 2020); ambiguity, long avoided in HCI, can be a resource that encourages close personal engagement (Gaver et al. 2003); making the familiar strange opens design spaces (Bell et al. 2005); visual difficulties can draw out active processing and aid comprehension and recall (Hullman et al. 2011); a measure that becomes a target ceases to be a good measure (Strathern 1997); and iteration that keeps refining one option gives rise to fixation (Dow et al. 2010). Among makers, colors can carry the message without numbers (Hawkins), hand-drawn charts invite viewers to question them (Chalabi), and data can be used to become more humane rather than more efficient (Lupi and Posavec). On the other hand, salient features of summary displays are misread as other quantities (Padilla et al. 2017), a wrong abstraction cannot be rescued downstream (Munzner 2009), and abstraction can also cut data off from people (Thorp). From these sources, the note proposes two hypotheses: the efficiency yardstick belongs to situations where the question is already fixed, and where it is not, a graphic becomes a place that generates questions rather than a tool that delivers answers; and the value of departing from efficiency arises only when the maker designs the departure (keeping accuracy and drawing the line between room for interpretation and misreading).

  • Where Does the Value of Showing Data as Graphics Lie? Kinds of Value and Why They Arise (from Industry Sources)

    2026-10-07

    This note examines the value of showing data as charts and visual displays in themselves, as distinct from data analysis or data science in general. It draws mainly on industry sources (the US FDA's experiment on front-of-package nutrition labels, the US FTC's research on appliance energy labels, the UK evaluation of smart meter in-home displays, the review of the US National Hurricane Center's forecast cone, the Japan Meteorological Agency's Kikikuru risk maps, interventions by the UK statistics regulator, figures published by Datawrapper, Spotify, and Apple, and statements by the makers of the warming stripes and Flatten the Curve), using academic research to supply the reasons. The value divides into immediate understanding (comprehension and speed), reach and sharing (communication and engagement), changed behavior, noticing abnormalities (safety and operations), continued use (product experience), and maintained trust. Among sources that compared graphics with other displays, nutrition labels with interpretive words helped people identify healthier foods better than labels showing numbers only (FDA, 9,200 participants), and 82% of people who saw an energy label showing efficiency in categories ranked refrigerators correctly by efficiency, compared with 72% for the current label. On the other hand, only 63% could read a scatterplot correctly and 21% a stock market chart, the forecast cone has been called the most misinterpreted graphic, and a study claiming that charts increase persuasiveness did not replicate. Effects on behavior are also small and conditional: a correlation of 0.071 for energy feedback and a 2.3% reduction in electricity use with in-home displays. From these sources, the note proposes two hypotheses: the value of a graphic arises less from its being a picture than from carrying benchmarks and evaluations in visual codes and thereby taking over the viewer's computation; and the value depends both on the maker's editing (what to cut and which benchmark to compare against) and on the reader's ability to read.

  • What Value Does Data Visualization Have in Industry? Market Valuation, Where Value Arises, and What Changes after Generative AI (from Industry Sources)

    2026-10-07

    This note examines the industrial and business value of data visualization (dashboards, business intelligence, news graphics, and the visualization of official statistics), drawing on industry sources: statutory filings and earnings calls of listed companies, records of the UK statistics regulator and health agency, Japan's Digital Agency and Cabinet Office, surveys by IDC Japan and BARC, the US Bureau of Labor Statistics and Indeed, and statements by the makers of visualization tools and news graphics. Academic research was not collected. The market places a high value on visualization. Salesforce acquired Tableau through an exchange of shares, Alphabet acquired Looker for $2.4 billion, and Power BI passed 30 million monthly active users. Meanwhile, the share of employees who actually use BI tools averaged 25% and had barely grown over seven years, and Domo, a stand-alone vendor, saw flat revenue while its net retention of revenue from existing customers fell from 106% to 89%. Value was clearly created where a single message was conveyed to the public, as with the UK COVID-19 dashboard (about 19 million page views a week, with trust in its data rising from 73.4% to 90.3%) and Our World in Data (from 2 million to 10 million monthly users). The makers place the value of a chart in the author's ability to state its point and in the context added to it. Inside organizations, they describe time consumed by maintaining dashboards and dashboards not being used for analysis. Since 2024, Copilot in Power BI and Tableau Pulse have added natural-language queries and notifications, but Microsoft itself writes that Copilot can return inaccurate answers when data are not prepared. From these sources, the note proposes three hypotheses: as charts become cheaper to produce, value moves upstream (preparing data and metrics) and downstream (editing messages and context); dashboards change from screens that people open to answers and notifications that reach them; and for charts aimed at the public, the value of showing sources and methods remains. No source in which a disinterested third party measured how much decision quality improved was found.

  • Drawing Networks as Data Art: Literature on Placement, Bundling, Time, and Flow

    2026-09-24

    A force-directed drawing of a network looks like the one true shape of that network, but it is only one of many arrangements. This note lays out the force equations of force-directed placement (Eades 1984, ForceAtlas2) and force-directed placement on a sphere (Kobourov and Wampler 2005), hive plots that set coordinates from attributes, edge bundling (Holten 2006), layouts along a single axis and along time (arc diagrams, ThemeRiver, streamgraphs, alluvial diagrams), mathematical marbling that makes patterns by dropping paint on water (Lu et al. 2012, Jaffer's drop mapping), prior works that treat data as art, and the curves and numerical methods used to draw the lines (centripetal Catmull–Rom curves, uniform B-splines, the midpoint method, Gaussian kernels, stream functions), and ways of rendering ink bleeding into paper (the watercolour model of Curtis et al. 1997, MoXi by Chu and Tai 2005, sumi painting by Guo and Kunii 1991), each with how far it was verified (full text, abstract, or bibliographic only). On this wiki's /network page, these methods were used to rearrange the same co-citation relations, which stay fixed while the arrangement changes with the method and the seed. The space each force-directed work is laid out in is fixed per work (marbling: a plane with the drawing area's aspect ratio; weave: an unbounded free plane; constellation: a sphere projected stereographically). The variation across seeds comes, for the weave, from initial-configuration families and ranges of force coefficients (including ForceAtlas2's LinLog and strong-gravity options), and for marbling and the constellation, from jitter in the initial placement and small ranges of coefficients.

  • What Data Visualization and Data Art Share, and Where They Differ

    2026-09-22

    The same line chart is called a chart in a newspaper and data art on a gallery wall. The two share their material, their visual variables, their makers, and even their venues, so the medium and the tools cannot separate them. Experimental work undercuts the common claim that visualization is about accuracy and art is about beauty, from both sides. Theories of art split three ways on whether the distinction can be settled objectively: by properties of the object, by no single criterion, or by institutions. Empirical studies show low agreement between judges, so no objective test on the object alone holds up. What can be recorded objectively is not which one a work is, but which one it was treated as, which makes the distinction usable as a procedure across three layers: what the figure asks of its reader, the intent and venue, and the institution that certified it.

  • How Data Visualization and Data Art Have Been Treated in Creativity Education

    2026-09-22

    Data education runs in two streams. One teaches students to read figures correctly and measures that ability with instruments (VLAT, CALVI, Mini-VLAT, a multidimensional scale), and it has accumulated critiques of those very instruments. The other brings the making of data works into the classroom through course design and public exhibitions, and it has no standard scale for outcomes. What bridges them is the making-to-learn lineage that descends from Papert's constructionism and the teaching of data physicalization. Arts-integrated practice reports the same tension again and again: artistic creativity against the accuracy of data. In the Japanese-language literature, research that uses data art in education is almost absent.

  • What Is Generative Art For in Education? Seven Purpose Types and a Literature Map of 72 Studies

    2026-08-02

    A literature map of 72 academic studies (86 explored, 14 merged as duplicates, 0 excluded) that use generative art (generative art / creative coding / computational art) as educational material, organized by the purpose of educational use. Seven purpose types emerge: (1) motivating and contextualizing introductory programming (media computation, Processing/p5.js); (2) inclusion of diverse learners (women, non-majors, low-income communities); (3) developing and assessing computational thinking (CT assessment of Scratch projects); (4) STEAM integration and constructionist making; (5) expanding expressive techniques in art, design, and music education; (6) creativity education; and (7) AI literacy and critical media literacy after LLMs (2023–2026). In terms of educational stages, CT development, STEAM, and inclusion dominate K-12; motivational introductions dominate university CS; and expressive expansion dominates art schools and higher education in architecture and music. Since 2023, 'education about writing code' has been rapidly joined by 'education using generative models as material for critical understanding' (one review reports that empirical studies of generative AI in art education grew from 2 in 2023 to 14 in 2025). All 72 items carry DOIs/URLs.

  • Adversarial Review of Two Generative Art Surveys: Examining the Literature Map and the Citation Strategy Against the Design-Theory Canon

    2026-08-02

    An adversarial examination of two notes — a generative art literature map (94 items) and a bibliometric analysis ranking citation-earning angles for a review article — through the design-theory lens (Simon/Schön/Cross/Dorst & Cross/Buchanan/Rittel & Webber/Krippendorff/Costanza-Chock). The literature map's four main weaknesses: an entanglement in which AI's ontological status (tool / co-designer / environment / evaluation apparatus) shifts from chapter to chapter; the unsorted concept of 'generation'; a sampling-inference leap that infers a disciplinary disconnect from absence in the 94-item corpus; and a compression of conceptual history that runs the generative Ästhetik of 1965 and the text-to-image of the 2020s through a single definition. The citation-strategy side's five main weaknesses: a slide from citation prediction to research value (a Schön-style critique of delegating problem-setting to the citation market); age bias in the citation-velocity metric and extrapolation to oneself; an exhaustiveness slide in which the qualified gap judgment loses its qualification in the No. 1 verdict; survivorship bias from looking only at highly cited surveys; and the implication leap of the 'cumulative 3,850-citation re-citation node.' At the same time, steelmanning established that the substance of the gap discovery (the unconnected state of computational-creativity evaluation and LLM empirical work) and the limited usefulness as circulation prediction cannot be overturned. All criticisms are tied to sources in an Evidence Ledger, and disagreements are left as a disagreement table rather than folded. The full critiques are stored in source/review/generative-art-adversarial-review/.

  • Which Angle for a Generative-Art Review Paper Attracts Citations? An Empirical Analysis Using Bibliographic Data

    2026-08-02

    This note empirically identifies, using public bibliographic data from OpenAlex and Semantic Scholar, the angles most likely to attract citations when writing a review paper on generative-art research. (1) Measured citations of 43 existing reviews show that 2023–2025 empirical studies (Zhou & Lee, 191.5 citations/year) and large technical surveys (Yang 2023, 288.2 citations/year) are fastest, while theoretical work holds a stable base of 14–30 citations per year. (2) In theme-level growth, human-AI co-creation is fastest at +607% from 2024→2025, followed by copyright at +160% and LLM creativity evaluation at +115%. (3) Highly cited surveys all include taxonomy and open-problems sections (7/7), and only the top-cited tier maintains GitHub repositories. Systematic reviews are cited more than other research designs (6.6 per year on average; JIF explains R²=0.59). (4) Verification of gap areas shows that a review connecting computational-creativity evaluation frameworks (Boden/SPECS/Lovelace) to the evaluation of LLM/diffusion-model outputs is most promising (a re-citation node for prior theory with over 3,850 cumulative citations, a 15.8-fold increase in primary studies, and the nearest review explicitly stating the connection is absent). The label-effect meta-analysis is no longer a gap, as De Rooij 2025 has already been published. Integrating these, the note ranks the seven most citation-promising angles. All figures carry sources and retrieval dates.

  • The Scholarly Lineage of Generative Art: A 94-Item Literature Map from Information Aesthetics to the Post-LLM Era

    2026-08-02

    A literature map of scholarly research on generative art, organized into four streams (94 items; 108 rows explored, 8 merged as duplicates, 0 excluded) corresponding to the chapters of a review article. (1) Historical origins: the lineage from Bense's and Moles's information aesthetics as the theoretical foundation, through the coining of generative Ästhetik at the 1965 Nees exhibition and the 'rot' magazine, the 1968 Cybernetic Serendipity and Zagreb New Tendencies, to the art world's rejection (Taylor) and the definitional debate (Galanter's autonomous-system definition). (2) Technical genealogy: the generational succession from L-systems and evolutionary computation (Sims, IEC) through GAN/CAN, StyleGAN, and diffusion models (DDPM, CLIP, Latent Diffusion) to LLM multimodal generation. (3) Theoretical frameworks: Boden's three types and their formalization (Wiggins, Jordanous's SPECS), the authorship debate (Hertzmann's tool argument versus McCormack's four concepts), and empirical reception studies showing that AI labels lower evaluations (Ragot, Bellaiche). (4) The post-LLM era (2024–2026, the thickest chapter): empirical text-to-image studies (AI adoption raises productivity +25% while average novelty declines), LLM code generation for creative coding (Spellburst, GenP5), human-AI co-creation, the arms race between copyright-protection tools and circumvention (Glaze/Nightshade vs. bypass attacks), and split empirical findings on labor effects (no short-term income decline vs. five-year longitudinal reports of job loss). Identified gaps include the disconnection between computational-creativity evaluation frameworks and post-LLM empirical research.

  • Research Currents in AI and the Study of Art and Culture: Aesthetics, Media Studies, Computational Creativity (2024–2026)

    2026-07-19

    A synthesis of 16 scholarly works on the intersection of generative AI and the study of art and culture, across aesthetics, computational creativity, media/cultural studies, musicology, and critical cultural theory. It traces how authorship and aura are being redefined, how creativity evaluation is being questioned, and how the politics of representation and labour moved to the centre. 16 references.

  • The Generative Art Ecosystem: Industry Intelligence

    2026-06-28

    An industry research note organizing generative art's copyright landscape, market data, tools, and practitioner testimony across three source tiers