Shuichiro Ogawa
日本語

Analysis and works from the notes

Citation Network

The 4,227 sources cited by 142 notes, analyzed as a network and drawn as seven works.

Each note on this site ends with its references: papers and web pages. Here each source is a point, and each note is a single line that ties together all the sources it cites, so the picture shows where notes meet on the same sources.

It covers the published notes written between June 2026 and September 2026. The same data is drawn seven ways, and the findings, the method, and the raw data are published alongside.

Overview

  1. 01

    Works

    The same data, drawn as three live works and four stills.

  2. 02

    Findings

    Four points on groups and periods, the source that links groups, and how notes overlap.

  3. 03

    Method

    What is counted from the notes and how it becomes a picture; the groups, the periods, the numbers, the raw data, and the references.

01

Works

The live works (weave, constellation, rings) work out where each source goes by forces, in your browser, every time the page opens. The stills (bundle, arcs, river, streamlines) draw the finished picture at once: the bundle, arcs, and river place things in a fixed order, and the streamlines use the positions from the same force computation as the live works.

Each thread is one note, and each knot where threads meet is a source cited by two or more notes.

Hover for a name; click for a card that links to the note or source. Press Esc or click an empty spot to close it. In the live works, “Generate another arrangement” lays the sources out again, and “Save image” saves the current picture.

Built with TypeScript and p5.js 2.3 (p5 runs the canvas and the draw loop; the shapes are drawn straight to Canvas 2D). The counting and grouping run at build time with graphology (the Louvain method and betweenness centrality); the force computation and the drawing of each work are custom code. The site is generated statically with Astro.

How this work is made (Weave)

Data used

  • Citation count
  • Co-citation count
  • Date written and period
  • Group
  • Kind

See what each quantity means ↓

What you see, and what it shows
What you see What it shows
Knot A source (cited by two or more notes)
Knot size Citation count
Filled / ring Kind (filled is a paper, a ring is a web page)
Knot shade Group (lighter is more web)
How close knots sit Co-citation count (the more, the closer)
Thread A note (it runs through the sources it cites, nearest first)
Dots around a thread Sources only that note cites (their positions mean nothing)
Small looped thread A note that cites no source shared by two or more notes
Order threads appear (period at lower left) Date written and period (oldest note first)
Red What you selected, and the most cited source

How it is made

Sources are tied by springs as strong as their co-citation count, every source pushes every other away, and each settles where these forces balance. Notes join one at a time in the order they were written; each adds a thread, its new sources, and their springs. The start changes every time the page opens, so the shape changes, but which sources sit near which stays much the same.

closeness ← co-citation count, weight ← citation count

Based on

Jacomy et al. (2014), Eades (1984), Fruchterman & Reingold (1991), Small (1973), Yuksel et al. (2009)

Built with

custom force simulation, and thread curves through the knots (centripetal Catmull-Rom). src/lib/weave-works/weave.ts, src/lib/weave-works/physics.ts

02

Findings

The sources fall into 11 groups (sources that tend to be cited together), and the notes fall into six periods by the day they were written.

  • Only “Watching the industry” is cited in all six periods.

    96 sources, 9% of them papers. Each of the other 10 groups goes uncited in at least one period.

  • Groups made almost entirely of papers are cited mostly in a single period.

    In each of the four groups where 95% or more of the sources are papers, at least 74% of the citations fall in one period.

  • The source that links groups most strongly is Doshi & Hauser 2024.

    It has the highest betweenness centrality (how often a source lies on the shortest paths between other sources; Freeman, 1977). The eight notes that cite it draw on sources from 7 of the 11 groups.

    Open Doshi & Hauser 2024 ↗

  • 524 pairs of notes share at least one source.

    That is 5% of the 10,011 possible pairs among 142 notes.

03

Method

Numbers are taken from the notes, and each work turns them into a picture by its own rules. The counting runs once, when the site is generated, so the same notes always give the same numbers.

  1. 01

    Notes

    142 published notes. Drafts are not counted.

  2. 02

    The list of sources

    DOIs and URLs are picked up from each note’s references, and repeats of the same source are merged: 4,227 sources.

  3. 03

    Count

    The six quantities below are counted.

  4. 04

    Draw by each work’s rules

    The quantities become position, size, shade, and lines. The rules are under each work, in “How this work is made”.

The six quantities

“How this work is made” under each work names the data it uses with these same words.

Citation count
How many notes cite the source.
Co-citation count
How many notes cite two sources together in the same note.
Known in analysis as co-citation (Small, 1973)
Shared-source count
How many sources two notes both cite.
Known in analysis as bibliographic coupling (Kessler, 1963)
Date written and period
The day a note was first written, and which of the six periods on the top page it falls in.
Group
A cluster of sources that tend to be cited together (11, found automatically), and the share of papers in it (more web or more academic).
Kind
Whether the source is a paper (DOI or preprint) or a web page.

In the weave, the constellation, the rings, the bundle, and the streamlines, only the 445 sources cited by two or more notes become points (knots), out of 4,227 sources cited. A source cited by only one note becomes one of the tiny dots around its thread in the weave.

The formulas and constants are in the analysis note, under “Formulas of the works”.

The groups

Sources that are often cited together are sorted into groups by the Louvain method (Blondel et al., 2008). How cleanly they separate is measured by modularity (Newman & Girvan, 2004). The calculation decides only which sources share a group; the names and descriptions were written by a person after reading what is inside.

The list runs from web-leaning (few papers) to academic-leaning, the same order as the shades in the pictures and the slices of the rings. A paper is one with a DOI or a preprint on arXiv and similar sites; a web page is a page whose content can change later.

  1. Public skill standards

    A national standard defining digital skills.

    0% papers

    1 source, cited most in period 5 (Mid-September 2026)

  2. Watching the industry

    Release notes of design tools, AI company announcements, market surveys and regulation pages. The places the weekly watch revisits.

    9% papers

    96 sources, cited most in period 1 (Early June to mid-July 2026)

  3. Delegating judgment to AI, and evaluation standards

    Design documents of an AI that returns judgments, and public documents on how to evaluate AI.

    25% papers

    51 sources, cited most in period 6 (Mid to late September 2026)

  4. Design systems

    Design system specifications, and studies checking whether they work.

    63% papers

    8 sources, cited most in period 5 (Mid-September 2026)

  5. LLMs as judges

    Using language models to grade and evaluate, and the limits of doing so.

    80% papers

    15 sources, cited most in period 2 (Mid to late July 2026)

  6. Research gaps and reframing problems

    Sources used to look for gaps in research and to rethink how a problem is framed.

    83% papers

    58 sources, cited most in period 2 (Mid to late July 2026)

  7. AI in design practice and education

    Studies of what happens when generative AI enters design work and design classrooms (mostly from the HCI conference CHI).

    90% papers

    52 sources, cited most in period 1 (Early June to mid-July 2026)

  8. Measuring dispositions and ability

    How to measure people's dispositions and abilities, and with what evidence.

    95% papers

    39 sources, cited most in period 3 (Early to late August 2026)

  9. Generative art and creativity

    Generative art, and research on how AI changes human creativity and the diversity of ideas.

    98% papers

    43 sources, cited most in period 3 (Early to late August 2026)

  10. How to frame research questions

    How to find new questions, and research through making (Research through Design).

    100% papers

    42 sources, cited most in period 2 (Mid to late July 2026)

  11. Learning and cognitive load

    Research on the mental load of learning and the supports that help people learn.

    100% papers

    40 sources, cited most in period 3 (Early to late August 2026)

The six periods

These are the same six periods as “Interests over time” on the top page. Time means when a note was first written (the first day it appears in the update log), not when it was last edited.

  1. 1

    Early June to mid-July 2026

    How AI changes designers' work and pay

    44 notes first written, 149 sources first cited

    Most cited group: Watching the industry

  2. 2

    Mid to late July 2026

    Research methods, and how questions are framed

    37 notes first written, 121 sources first cited

    Most cited group: Research gaps and reframing problems

  3. 3

    Early to late August 2026

    Learning, and measuring ability

    34 notes first written, 115 sources first cited

    Most cited group: Generative art and creativity

  4. 4

    Late August 2026

    AI slop, and where design as a field came from

    10 notes first written, 18 sources first cited

    Most cited group: Watching the industry

  5. 5

    Mid-September 2026

    Design systems, and the standards work sits on

    5 notes first written, 11 sources first cited

    Most cited group: Design systems

  6. 6

    Mid to late September 2026

    Judgments handed to AI, and thinking by making

    12 notes first written, 31 sources first cited

    Most cited group: Delegating judgment to AI, and evaluation standards

“Sources first cited” are sources cited by two or more notes that a note from that period cited first. The counts differ from the top page because only published notes are counted here, and the regular watch notes are included.

What position and distance mean
  • In the weave, the constellation, and the streamlines, where each source sits is not decided ahead of time; it sits where these forces balance. Sources with a higher co-citation count pull on each other harder. Every source pushes every other away. The whole is pulled gently toward the middle. A source with a higher citation count is heavier and pushes a little harder. The rings use similar forces, but each source moves only inside a fixed cell. The forms of the forces follow ForceAtlas2 (Jacomy et al., 2014); the weight was added for this site.
  • When the effect of weight was measured, a heavy source moved less for the same force, but in practice it moved about as far as a light one, because a source cited more often also has more partners pulling on it.
  • Only relative closeness means something: sources close together tend to be used in the same notes. The overall shape and direction (which way is up or right) change every time and mean nothing. Distances cannot be measured with a ruler and compared, and closeness in a force-directed picture always leaves room for interpretation (Venturini et al., 2021).
  • A force-directed picture depends on where the points start. Gansner et al. (2005) drew the same graphs from 25 random starts to compare them. Showing a single arrangement makes a chance shape look like the answer, so this page grows a different arrangement each time. The number next to the buttons (such as “Arrangement 1234”) is the seed of the arrangement, and it is also added to the URL after the #. Opening a URL with the same number reproduces the same arrangement. Sources that stay close across several arrangements are the most strongly tied.
  • The group “Public skill standards” has a single source that is never cited together with another, so its position means nothing.
Numbers
Published notes
142
Sources cited
4,227
2,089 with a DOI (an identifier for a paper)
Sources cited by 2+ notes
445
knots in the weave, stars in the constellation
Pairs of sources cited together
5,870
the lines of the constellation
Groups
11
sources that tend to be cited together
Modularity
0.73
how cleanly the groups separate; closer to 1 is cleaner

137 notes had at least one readable source. The 33 unpublished draft notes are not counted.

Raw data and analysis notes
  • Raw data (JSON)

    The same data as this page, rebuilt every time the site is generated.

  • Analysis note

    The steps and results of the calculation.

  • Literature note

    Prior work on network visualization and data art methods.

References

The analysis and the seven works stand on the methods and ideas below. Mapping fields through citations has been surveyed as the visualization of knowledge domains (Börner et al., 2003), and CiteSpace, described by Chen (2006), is a tool built for it.

What each work is built on

Weave
Jacomy et al. (2014), Eades (1984), Fruchterman & Reingold (1991), Small (1973), Yuksel et al. (2009)
Constellation
Jacomy et al. (2014), Eades (1984), Fruchterman & Reingold (1991), Small (1973)
Rings
Jacomy et al. (2014), Eades (1984), Fruchterman & Reingold (1991), Small (1973), Krzywinski et al. (2012)
Bundle
Holten (2006), Holten & van Wijk (2009), de Boor (1978)
Arcs
Wattenberg (2002), Kessler (1963)
River
Byron & Wattenberg (2008), Havre et al. (2002), Silverman (1986)
Streamlines
Jobard & Lefer (1997), Turk & Banks (1996), Batchelor (1967/2000)

Methods of the analysis

Placing by forces (weave, constellation, rings)

Algorithms for the stills (bundle, arcs, river, streamlines)

  • Holten, D. (2006). Hierarchical edge bundles: Visualization of adjacency relations in hierarchical data. IEEE Transactions on Visualization and Computer Graphics, 12(5), 741–748. https://doi.org/10.1109/TVCG.2006.147
  • Holten, D., & van Wijk, J. J. (2009). Force-directed edge bundling for graph visualization. Computer Graphics Forum, 28(3), 983–990. https://doi.org/10.1111/j.1467-8659.2009.01450.x
  • Wattenberg, M. (2002). Arc diagrams: Visualizing structure in strings. In IEEE Symposium on Information Visualization (InfoVis 2002) (pp. 110–116). IEEE. https://doi.org/10.1109/INFVIS.2002.1173155
  • Byron, L., & Wattenberg, M. (2008). Stacked graphs: Geometry & aesthetics. IEEE Transactions on Visualization and Computer Graphics, 14(6), 1245–1252. https://doi.org/10.1109/TVCG.2008.166
  • Havre, S., Hetzler, E., Whitney, P., & Nowell, L. (2002). ThemeRiver: Visualizing thematic changes in large document collections. IEEE Transactions on Visualization and Computer Graphics, 8(1), 9–20. https://doi.org/10.1109/2945.981848
  • Jobard, B., & Lefer, W. (1997). Creating evenly-spaced streamlines of arbitrary density. In Visualization in Scientific Computing ’97 (pp. 43–55). Springer. https://doi.org/10.1007/978-3-7091-6876-9_5
  • Turk, G., & Banks, D. (1996). Image-guided streamline placement. In Proceedings of the 23rd Annual Conference on Computer Graphics and Interactive Techniques (SIGGRAPH ’96) (pp. 453–460). ACM. https://doi.org/10.1145/237170.237285

Curves and numerical methods

  • Catmull, E., & Rom, R. (1974). A class of local interpolating splines. In R. E. Barnhill & R. F. Riesenfeld (Eds.), Computer aided geometric design (pp. 317–326). Academic Press. https://doi.org/10.1016/B978-0-12-079050-0.50020-5
  • Barry, P. J., & Goldman, R. N. (1988). A recursive evaluation algorithm for a class of Catmull-Rom splines. In Proceedings of the 15th Annual Conference on Computer Graphics and Interactive Techniques (SIGGRAPH ’88) (pp. 199–204). ACM. https://doi.org/10.1145/54852.378511
  • Yuksel, C., Schaefer, S., & Keyser, J. (2009). On the parameterization of Catmull-Rom curves. In 2009 SIAM/ACM Joint Conference on Geometric and Physical Modeling (pp. 47–53). ACM. https://doi.org/10.1145/1629255.1629262
  • Yuksel, C., Schaefer, S., & Keyser, J. (2011). Parameterization and applications of Catmull–Rom curves. Computer-Aided Design, 43(7), 747–755. https://doi.org/10.1016/j.cad.2010.08.008
  • de Boor, C. (1978). A practical guide to splines. Springer. ISBN 978-0-387-90356-9
  • Bartels, R. H., Beatty, J. C., & Barsky, B. A. (1987). An introduction to splines for use in computer graphics and geometric modeling. Morgan Kaufmann. ISBN 0-934613-27-3
  • Press, W. H., Teukolsky, S. A., Vetterling, W. T., & Flannery, B. P. (1992). Numerical recipes in C (2nd ed.). Cambridge University Press. ISBN 0-521-43108-5
  • Silverman, B. W. (1986). Density estimation for statistics and data analysis. Chapman & Hall. ISBN 0-412-24620-1
  • Batchelor, G. K. (2000). An introduction to fluid dynamics. Cambridge University Press. (Original work published 1967) https://doi.org/10.1017/CBO9780511800955 The link is the DOI of the Cambridge Mathematical Library edition (2000).

Precedents in art

  • Lima, M. (2011). Visual complexity: Mapping patterns of information. Princeton Architectural Press. ISBN 978-1-56898-936-5
  • Lupi, G., & Posavec, S. (2016). Dear data. Princeton Architectural Press. ISBN 978-1-61689-532-7
  • Posavec, S. (n.d.). Writing without words [Artwork]. Stefanie Posavec. https://www.stefanieposavec.com/archive/writing-without-words
  • Lombardi, M. (n.d.). Narrative structures [Drawings]. Whitney Museum of American Art. https://whitney.org/artists/7340
  • Viégas, F. B., & Wattenberg, M. (2007). Artistic data visualization: Beyond visual analytics. In Online Communities and Social Computing (OCSC 2007) (Lecture Notes in Computer Science, pp. 182–191). Springer. https://doi.org/10.1007/978-3-540-73257-0_21
  • Kosara, R. (2007). Visualization criticism: The missing link between information visualization and art. In Proceedings of the 11th International Conference Information Visualization (IV ’07) (pp. 631–636). IEEE. https://doi.org/10.1109/IV.2007.130
  • Viégas, F., & Wattenberg, M. (2012). Wind map [Interactive artwork]. hint.fm. http://hint.fm/wind/ Launch year (2012) and authors checked on the official page (it does not open over https, so the http address is given).