Lane Harrison

We do 8 weeks! (For undergrad at least…)

What are your favorite examples for people using #datavis to tell about aspects/parts/events of their lives?

(would watch a 30-minute news segment on this sort of analysis...)

Screenshot of paper titled “A Priest, a Rabbi, and an Atheist Walk Into an Error Bar: Religious Meditations on Uncertainty Visualization” by Michael Correll and Lane Harrison. Below the title, an eye in a winged triangle is encircled by shapes and a ribbon with religious symbols. Text at the bottom reads “BE NOT CERTAIN.”

Find out what happens when “A Priest, a Rabbi, and an Atheist Walk Into an Error Bar 🍸”. Join @birdbassador.jorts.horse.ap.brid.gy & @laneh.bsky.social for their religious meditations on uncertainty visualization at alt.VIS, Sunday 9 AM - 12:30 PM!
#ieeevis #uncertaintyvis
arxiv.org/abs/2509.08213

Home | ReVISit reVISit: Reproducible and Powerful Visualization User Studies

Revisit by @laneh.bsky.social and team is an excellent option. It requires some configuration and setup, but it works beautifully and all your data stays on your server. (revisit.dev)

here's a preprint of the unhinged paper i wrote with @lane

we stop just short of saying that the unthinking use of frequentist statistics puts your very soul at hazard
https://arxiv.org/pdf/2509.08213

The reVISit team will be hosting an open hackathon at TU Graz in Austria, immediately after IEEE VIS from Monday November 10 to Wednesday November 13. Anyone interested in contributing to reVISit is invited to participate!

@hrbrmstr.dev thanks for the ping!

@erickapitanski.bsky.social very cool project (poking around now) -- I'll be in CA in a few weeks to advise student projects, would be happy to swing by if doable

dithered hypercard-style picture of a winged all-seeing eye being orbited by bar charts with cross error bars, flanked by discrete gaussian distributions

spent a chunk of today making biblically accurate error bars

Super love seeing regional datavis events making a comeback (and happy to be in town!) ->

So modest! Had an AVD-inspired hour+ long convo just the other week 😁

A screenshot of the reVISit interface showing the answers and provenance from a participant who completed the question. The study shown is a replication of He et al.’s pattern design study.

Going to #CHI2025 in Japan? Interested in running user studies?

Sign up for our virtual CHI 2025 course that teaches you how to use reVISit, a new, open-source user study tool reducing the tedium of study design. Come learn through some practical examples.

cvent.me/g5mx2w
revisit.dev

A new way to make graphs more accessible to blind and low-vision readers Designed to help blind and low-vision readers understand graphics, the “Tactile Vega-Lite” system from MIT CSAIL converts data into a standard visual graph and a tactile chart. Accessibility standards...

Tactile charts are an important tool for conveying data to blind and low vision people via embossed paper. However, tactile charts require high levels of time and expertise to design.

Introducing Tactile Vega-Lite, new work at #CHI2025 led by MIT SM student Katie Chen

news.mit.edu/2025/making-...

behold the CONNECTED SPLATTERPIE

kneel before my works, ye mighty, and despair

AI-generated parody of Minard‘s famous visualization of Napoleon‘s march to Moscow, embellished with all kinds of Studio Ghibli-inspired characters

The future of #datavis

1) Cool VIS study, 2) Help our replication effort! ->

Also curious about what might be driving this. Are folks with less experience diving into more complex tool ecosystems thanks to AI? Kind of what @mjskay.com was getting at... 🤔

Pre-AI, maybe folks learned better norms for asking for help, etc. over time?

design constraints good

A chart of country noises (lawnmower, rain on roof, frog, mosquito and more) presented to look like sound waves whereas they are more like ornamental border patterns - especially like old letterpress and printing pieces. This is probably what was used for them.

Country Noises.

(Saul Steinberg, 1979)
saulsteinbergfoundation.org

There be dragons: dataviz in the industry I started being serious about data visualization around 2005, when the field was still pretty niche and people like Martin Wattenberg, Ben…

"…there’s one nut we have not been able to crack: anchoring bespoke data visualization as a full fledged profession in corporate settings."

Re-reading this article from 2017 (medium.com/visualizing-...), I wonder — where are we at 8 years on?

Slightly annoying when a paper describing an experiment doesn't include a screenshot of the actual task participants performed.

(I wrote the paper in question 🫠...)

Cool possibilities for interaction studies ->

This ggplot chart was generated without any theming support, *at all*.

Image or Information? Examining the Nature and Impact of Visualization Perceptual Classification How do people internalize visualizations: as images or information? In this study, we investigate the nature of internalization for visualizations (i.e., how the mind encodes visualizations in memory)...

💯 + reminds me of this recent VIS paper "Image or Information" -> arxiv.org/abs/2307.10571

Any freelance Svelte/data visualization developers among my followers? Looking for someone with some availability, starting in a couple of months

revisitpy-examples/example_jnd_study/example_jnd_study.ipynb at main · revisit-studies/revisitpy-examples Example directory for all revisit-py features. Contribute to revisit-studies/revisitpy-examples development by creating an account on GitHub.

See it in action in the clip ⬆️, where we replicate a correlation perception experiment in ~160 lines of Python.

💻 Code available here ➡️ github.com/revisit-stud...

🛫 You can pilot and iterate on the experiment instantly from within a notebook environment.

📈 You can even test data collection and prototype your analyses in the same notebook!

ReVISit 2.0 (revisit.dev) reached several exciting milestones recently:

📁 You can now define an entire experiment in a single file— from trial data, stimuli (visualizations and interactions, thanks to Vega-Altair), study blocks/sequences and randomization, questionnaires, etc.

Crowdsourced Think-aloud Studies w/ full interaction provenance+rehydration, auto voice transcription, and correlation between timelines.

That plus a pile of new features and capabilities in the reVISit 2.0 release.

Check it out & let us know what you think!

More LLM reflections

I think a useful model is "everything that in-house counsel is not"

In-house counsel doesn't know how to say yes to you; they're entirely made of blood-soaked wisdom.

LLMs are excitable improv comedy partners with more memory than wisdom; they don't know how to say no to you.