Professor at CMU HCII (where I co-direct @cmudig) and ML researcher at ๏ฃฟ. I build visualization tools and systems for interactive analysis for everyone.
First was a thought-provoking talk by @dom on the importance for HCI researchers to build more interoperable, focused systems at Stanford University https://www.youtube.com/watch?v=6DIIkX19ihs (2/5) #HCI
Very excited for all of the nice things Foundation Model access unlocks in iOS 26. Makes things like automatic tag suggestion in Crouton super smooth.
@dave_andersen Home assistant is so much fun and so flexible. Nice application!
Here is my #chi2025 paper explorer. Visually explore papers by similarity, search for an specific topic or even find what specific sessions to attend. Please let me know what you think! #dataviz @chi
๐๐ผ https://johnguerra.co/viz/chi2025Papers/ ๐๐ผ
Coworker: ...and the IP address are compared with a string match.
Me: grinning manically
Coworker: Why are you looking at me like that?
Me: Open up a terminal and type ping 4.2.514
and hit enter.
Coworker: ...what's the fourth number?
Me: grin widens Just hit enter.
Coworker: WTF!?
Iโm at @chi ๐ธ next week so come and say hi ๐.
My group, the @cmudig, has a few presentation you donโt want to miss https://dig.cmu.edu/2025/03/21/chi2025.html.
We will also have awesome folks from Apple ๐ in Yokohama. Check out https://machinelearning.apple.com/updates/apple-at-chi-2025.
New blog post out (finally) โจ I tested the dataviz accessibility of the 2024 presidential election dashboards, and (as expected) found a lot of issues to write about:
https://fossheim.io/writing/posts/2024-dataviz-a11y-elections/
@pg I'd like more!
This is a great example of why the web rocks. Safari and Chrome disagreed on the spec, so we ended up with a third, better spec. This wouldn't happen if one company controlled everything: https://webkit.org/blog/16587/item-flow-part-1-a-new-unified-concept-for-layout/
@pg community engagement like sponsoring a conference or hiring someone would be the biggest expenses. Server costs or buying energy drinks are peanuts.
Super excited that Vega is now a NumFocus sponsored project. Announcement at https://bsky.app/profile/vega-vis.bsky.social/post/3lkqhrlro6c2y.
This New Open-Source Alternative to Google Docs and Notion Is Backed by France and Germany
Docs is an web-based collaborative text editor that is the result of a joint effort led by the French ๐ซ๐ท๐ฅ (DINUM) and German ๐ฉ๐ช๐ฅจ governments (ZenDiS).
It's a good alternative to popular services like Notion, Google Docs and Outline.
The platform offers real-time collaboration capabilities, a sleek minimalist interface, and all the essential features you could possibly need.
Best of all, Docs is open-source on GitHub and can be self-hosted, making it a viable option for local agencies and companies.
You can test Docs on your browser by visiting this page:
https://impress-preprod.beta.numerique.gouv.fr/
using the login:
email: test.docs@yopmail.com
password: I'd<3ToTestDocs
https://www.howtogeek.com/docs-alternative-google-docs-notion-france-germany/
You can also volunteer to help translating Docs to other languages here: https://crowdin.com/project/lasuite-docs
#europe #europeanalternatives #germany #france #opensource #BuyFromEU
advice for students --
as much as it is important not to uncritically accept AI hype, claims of superhuman performance, AGI, etc.
it is also important not to uncritically accept that LLMs are useless b/c they are sometimes wrong, that all AI is terrible for the environment, etc.
stay rigorous folks!
And finally, take a first look at HTML export: With it, you can export a website from the same source code as your PDFs. HTML export is currently incomplete and experimental. It is only available on the command line, behind a feature flag.
@hrbrmstr Yaaay, super exciting. And hopefully it's getting proper tagged PDF support soon as well to be something I feel more comfortable recommending.
Do you work with text data for your job or research? Join our user study!
โ Who? Programmers, researchers, or anyone working with text-based datasets
๐ What? 2โ3 sessions where youโll share how you work with text data and try out our prototype tool (up to $60 compensation)
๐ Data? Any text datasetโsocial media, news, research abstracts, etc.
Your feedback will help our research and shape our open-source visualization tool.
Learn more & sign up here ๐ https://forms.gle/beUBzERwiPQ9iVLEA
A reminder to those who manage #websites (I needed it myself):
Remove the share to X/Twitter button from your sites, if you no longer want to be associated in any way to a social network run by a billionaire that thinks it ok to do a Nazi salute.
ah, #chi2025, where the decisions are random but impactful. as always happy to be faculty submitting enough papers where the statistics mean i got some accepted.
but, sad esp that one really nice paper, that went into R&R w/ good reviews, to which a lot of work was done, was rejected.
anyway, you should check out "AI Policy Projector: Grounding LLM Policy Design in Iterative Mapmaking",
https://arxiv.org/abs/2409.18203
What is the best/most fun snow sculptures you made/seen?
I will put them in a (new) post on streetartutopia.com