We design for people like us. That's the problem. New 4-part series on using (AI-assisted) personas to see your work through someone else's eyes.
ai.ericoosenbrug.com
Feedback from public servants very welcome.
#AI #PublicService #GovTech
Wrapped up a 4-part series on validating AI-generated analysis for public servants. Practical, not theoretical.
ai.ericoosenbrug.com
Early days for this blog β feedback from public servants very welcome.
#AI #PublicService #GovTech
Launched 'Thinking Through AI' - a blog on effective AI use in public service. Latest: 4-part series on validating AI-generated data analysis ('test first, implement second' technique actually works). ai.ericoosenbrug.com #OpenGov
Loved going on Cam Newport's show for an episode called "Was 2025 A Good Or Bad Year for AI?" and man, is that a complex question. One of my favourite interviews.
www.youtube.com/watch?v=gJ8p...
βThe point of modern propaganda isn't only to misinform or push an agenda.
It is to exhaust your critical thinking, to annihilate truth.β
- Garry Kasparov
When I first studied statistics in high school, it appeared to me not as a revelation, but as a jumble of bizarre formulas that defied explanation. I could not for the life of me make sense of it. The calculations were mechanical and joyless. Compared to the elegance of physics, these equations felt unappealing, artificial, and just plain wrong. There was no reasoning to follow, only rules to memorize. I decided with absolute finality: statistics was not for me.
Read it here: open.substack.com/pub/kareemca...
Spent time with #Tableau calculated fields today. Created Product Value Category field & ran into problem - some products categorized as BOTH categories. Took a min to realize: transaction-level data means same product, different prices. Note to self: check data structure before building categories.
Getting more confident with calculated fields in Tableau. Understanding how to create categories and conditions without modifying source data. Game-changer for analysis flexibility. Small wins add up. #datafam What Tableau feature clicked for you once you actually used it?" #dataviz
Spent 2 hours yesterday working on Tableau fundamentals. Still very much a beginner learning how to work with calculated fields, but I'm committed to building real fluency with data tools. #datafam For anyone else learning #dataviz: what's a tricky technical concept you're wrestling with right now?
Venn diagrams as jack-o-laterns showing trick/treat logicals: OR, AND, XOR, NOR, NAND, and XNOR
Halloween logicals, still the best 10/31 venn
this guy (and this @wired.com series) is excellent youtu.be/YvZPU-KBCiQ?... π¦
One of the loudest bells tolling for social science right now is that the decade of abundant data on humans is coming to a close. Whether you work on digital trace or surveys, LLM pollution is a serious problem, even while the wide roll out of LLMs creates an urgent need for social science.
Re-upping this again. I assisted the amazing team at @onthemedia.bsky.social developing the breaking news consumers handbook, and its recommendations still hold true to this day.
Is MAGA a community or is it a cult? π€·
youtu.be/86tcoDjJWbk?... I agree using pejorative labeling is weaksauce, but also true here? π
The biological transhumanism and digital eugenics promoted by some of the most powerful people in Silicon Valley amounts to pro-extinctionism, @xriskology.bsky.social⬠writes. www.techpolicy.press/digital-euge...
When Trump demands foreign countries pay tariffs, they're laughing at him π€‘ No country can force another to pay its taxes π If China imposed taxes on Americans - they'd simply ignore them. US importers pay tariffs, American consumers end up footing the bill #EconomicsExplained #TrumpTariffs π«‘
Graph theory shows up where you least expect it. The mathematician Maria Chudnovsky even used it to assemble the seating chart at her wedding. Tune in to @jannalevinastro.bsky.social βs conversation with her on this weekβs episode of βThe Joy of Whyβ: www.quantamagazine.org/how-does-gra...
Chatbots β LLMs β do not know facts and are not designed to be able to accurately answer factual questions. They are designed to find and mimic patterns of words, probabilistically. When theyβre βrightβ itβs because correct things are often written down, so those patterns are frequent. Thatβs all.
Ξ£ Waterfall charts are perfect for showing both individual contributions & cumulative impact. Each bar shows how much each subcategory adds to total- you can instantly see which categories drive the biggest changes. Easy in #Tableau with calculated field (sales difference) #dataviz #LearningInPublic
β‘ Arrays and matrices are basically the same (unlike in math/programming) because #Tableau treats everything as "rows Γ columns." Arrays store lists of values, matrices represent data structure - but both are just rectangular grids. I was probably just overthinking this π #dataviz #LearningInPublic
β Until recently, I sometimes confused dimensions vs attributes in #Tableau (because they're both categorical) - classic case of surface similarity bias. But now I know dimensions structure your viz (create groups), while attributes just add context (annotate). #dataviz #datapsych #LearningInPublic
π€― #Tableau hierarchy order matters! I thought they were just groupings but they're drill-down paths (duh). My "database table" mindset made me miss the tree structure hiding in plain sight π§
#dataviz #datapsych #LearningInPublic
I resonate with the feeling/need to connect at conferences. I'm new to dataviz and decided not to go ... I regret it. I wish I had met new people and made some new friends. Appreciate the share!
The second edition (on Amazon) is listed as "Volume 1" - is there another volume coming? Just curious
This has absolutely happened to me. It's a good self-reminded to sloooow down :)
This is so helpful Whitney, thank you for sharing! I appreciate you providing these kinds of tips to us data n00bs :)
This looks fantastic and timely! What an amazing line up of researchers. Can't wait to read these essays π
Awesome work! I'm so surprised there has never been an Indian dataviz conference - such an important endeavor - best of luck!
Documenting my journey from psych PhD to data storyteller, imposter syndrome and all.
This week: I'm wrestling with basics of data modeling & ETL - guess I missed part about making data usable in my stats classes π
What's do you wish someone had told you when you started #dataviz? #LearningInPublic
Iβm standing next to a woman and weβre holding her sign together: βSTEPHEN MILLER IS A LITTLE BITCH!β
Iβm Kat Abughazaleh and I approve this message.
