Lightnews — Scholar-powered news

Matthew Kay

@mjskay.com

11K followers 1K following 1.6K posts

Assoc Prof Computer Science and Communication Studies at Northwestern. Infovis, HCI. Author of tidybayes & ggdist R pkgs. he/him. 🏳️‍🌈 https://mjskay.com/ Co-director https://mucollective.northwestern.edu Co-founder https://journalovi.org

mjskay.com

Posts Media Videos Starter Packs

Pinned

Matthew Kay @mjskay.com · Sep 7

I work in #datavis, best known for uncertainty vis: see my work on election forecast vis (forecasts.cs.northwestern.edu), this talk (youtu.be/E1kSnWvqCw0), or my lab (mucollective.northwestern.edu)

I like writing #rstats 📦s: ggdist (mjskay.github.io/ggdist), tidybayes, ggblend, posterior::rvar...

Some examples from the three main families of ggdist geometries: slabinterval, dotsinterval, and lineribbon

An example of an election forecast communicated using a quantile dotplot, from: Yang, Fumeng, Chloe Rose Mortenson, Erik Nisbet, Nicholas Diakopoulos, and Matthew Kay. "In Dice We Trust: Uncertainty Displays for Maintaining Trust in Election Forecasts Over Time." In Proceedings of the CHI Conference on Human Factors in Computing Systems, pp. 1-24. 2024.

3 12 81

Matthew Kay @mjskay.com · 2h

Good APIs typically have some clear structure and rules - relational algebra (dplyr), grammar of graphics (ggplot) - around which a consistent API can be built

Several parts of base R also (locally) have this property, but globally the API is less coherent (the price of backwards incompatibility?)

1 2

Matthew Kay @mjskay.com · 2h

This is nice! I like the observation that unlike traditional user-centered design, the tidyverse (and really most APIs in general) don't start from studying users but from studying the domain itself (though package developers are often also their own users).

1 2

Reposted by Matthew Kay

Emi Tanaka @emitanaka.org · 2h

I wrote a paper about "Examining the Interface Design of Tidyverse" 📝

TL;DR to follow in future 😊

arxiv.org/abs/2510.10382

Examining the Interface Design of Tidyverse

The tidyverse is a popular meta-package comprising several core R packages to aid in various data science tasks, including data import, manipulation and visualisation. Although functionalities offered...

arxiv.org

2 2 11

Matthew Kay @mjskay.com · 12h

lol welp

Matthew Kay @mjskay.com · 13h

also makes it more legible to drivers, which presumably is the audience? anyway

Matthew Kay @mjskay.com · 13h

It's also like, 5 extra minutes of thought solves the problem?

e.g. flip it so the baseline is where the cars stop and for extra measure fill in above the bars with something (different pattern or color)

1 3

Matthew Kay @mjskay.com · 13h

I suppose there's always been some version of this (for me it's often been "kids these days don't even know what a relational database is") but it is exacerbated by a technique that can really quickly do a 75%-okay solution on lots of things but which has no path to get to 99%

Matthew Kay @mjskay.com · 13h

especially frustrating for longstanding problems with well-known, reliable solutions

2 6

Reposted by Matthew Kay

Susan Vanderplas @srvanderplas.bsky.social · 3d

Today is the day we defend the #stats department against the #unl #administration. If you haven't yet left a comment apc.unl.edu/fall-2025-bu..., please do so and explain how silly it is to cut #statistics in this day and age. Hear us make our case at 3 Central today: unl.zoom.us/j/99178278682.

Fall 2025 Budget Reduction Feedback Form | Academic Planning Committee | Nebraska

apc.unl.edu

1 1 2

Matthew Kay @mjskay.com · 3d

yes except who says I hate fine-tuning a plot

1 8

Matthew Kay @mjskay.com · 4d

would you say you're more a 75° or a 10°

1 4

Matthew Kay @mjskay.com · 4d

ohoho I used to follow a typography blog around when that pepsi rebrand happened and people just tore. them. apart.

Matthew Kay @mjskay.com · 4d

omg I forgot about time cube!!! truly a golden era

1 1

Matthew Kay @mjskay.com · 4d

though this is more "can you eyeball about 10% of the mass being above water and the center of mass being in the middle so the thing doesn't move from initial conditions" which obviously doesn't tell you if it could have *formed* that way in the first place

an apparent iceberg above water and a bunch of squiggles under the water

1 2

Matthew Kay @mjskay.com · 4d

and also

(this is fun I'm getting pretty good at this)

another iceberg, looks more like the original

1 3

Matthew Kay @mjskay.com · 4d

okay but

iceberger game with an iceberg that's got a long vertical pointy bit under the water

1 3

Reposted by Matthew Kay

Matthew Varona @matthew.wiki · 5d

Unreasonably excited to present THEORY IS SHAPES at alt.vis this November! Read on to find out what the "BLT Sandwich Theory of Visualization Consumption" (and other shape-based shenanigans) can teach us about theorycrafting in research 🥪🧲🧊♾️

Matthew Kay @mjskay.com · 5d

to appear at #alt.vis 2025: THEORY IS SHAPES

sure "theory figures" are great, but @matthew.wiki, @maryamhed.bsky.social, me, and Carolina Nobre wonder: why always a 2D plane or a flowchart? why not icebergs, horseshoes, Möbius strips, or BLT sandwiches?
arxiv.org/abs/2510.01382

#ieeevis #hci

Abstract and teaser figure of the paper "Theory is Shapes" by Matthew Varona, Maryam Hedayati, Matthew Kay, and Carolina Nobre.

The teaser figure is an "iceberg" theory figure, showing four levels: the tip of the iceberg, containing a nested set diagram and a Cartesian plane. The next level down (just under the water) has a complicated flow charts and a combination flow chart and matrix. The next level down has the iceberg figure itself and a horseshoe. The final level contains a Möbius strip and a BLT sandwich.

The abstract reads:

"Theory figures" are a staple of theoretical visualization research. Common shapes such as Cartesian planes and flowcharts can be used not only to explain conceptual contributions, but to think through and refine the contribution itself. Yet, theory figures tend to be limited to a set of standard shapes, limiting the creative and expressive potential of visualization theory. In this work, we explore how the shapes used in theory figures afford different understandings and explanations of their underlying phenomena. We speculate on the value of visualizing theories using more expressive configurations, such as icebergs, horseshoes, Möbius strips, and BLT sandwiches. By reflecting on figure-making's generative role in the practice of theorizing, we conclude that theory is, in fact, shapes.

1 5 8

Matthew Kay @mjskay.com · 5d

also I gotta say @matthew.wiki and @maryamhed.bsky.social did a great job with this --- it's hilarious

(I can say this without tooting my own horn because with the exception of a single half sentence my direct contributions were in feedback not writing)

Matthew Kay @mjskay.com · 5d

4 3 20

Matthew Kay @mjskay.com · 5d

The Confounders would 100% have started as a brother/sister twee-ish aughts indie folk band (think Iron & Wine meets Fiery Furnaces) and over decades shifted through baroque pop, post-rock, and 80s new wave revival to land on some strange edm mashup of like Kraftwerk and Godspeed You! Black Emperor

Matthew Kay @mjskay.com · 6d

good call! or even "insquariance"?

hmm I see definite pedagogical advantages to squariance / rootsquariance / insquariance...

Matthew Kay @mjskay.com · 6d

missed opportunity to rename all dispersion measures following the same scheme

OG variance = squariance
SD = stairiance (also known as rootsquariance)
Gini = fairiance
IQR = iquariance (pronounced with a schwa, like "aquarius")
MAD = marriance
precision = prairiance
etc

1 13

Matthew Kay @mjskay.com · 7d

ah yes, pull request check, perfect 👍

1 4

Reposted by Matthew Kay

Evan Peck @peck.phd · 10d

omg - this alt.vis paper about NASA-TLX was styled like the original NASA-TLX paper 😍

arxiv.org/pdf/2509.24643

arxiv.org

1 2 17

Matthew Kay @mjskay.com · 10d

don't sell yourself short! I see lots of small touches that elevate it: pleasing (and colorblind-safe) colors, densities are lighter shades of corresponding intervals, interval lines are just-thick-enough-but-not-too-thick, y axis labels are stacked to match subgroups (+ all text horizontal), ...

1 9