My article on uncertainty vis in SciAm is open for a few days! I hope it makes the research on this topic more accessible. My summary includes work by MU Collective (me & @mjskay) @LacePadilla @Birdbassador & @nytgraphics among others
Jessica Hullman
3,931 posts
Jessica Hullman
@JessicaHullman
Ginni Rometty Prof @NorthwesternCS | Fellow @IPRatNU
AI | uncertainty | decisions | metascience
Blog @statmodeling
Essays jessicahullman.substack.com
- Often we stop interpreting experiment results at an average treatment effect (ATE). But the same ATE can result from different patterns of variation. Causal quartets are like Anscombe's quartet, but for illustrating different patterns of variation we might see across units. 1/
with Lauren Kennedy and Jessica Hullman: “Causal quartets: Different ways to attain the same average treatment effect” statmodeling.stat.columbia.edu/2023/02/24/cau… - It's been a bit under a year since I started this paper w/@StatModeling, now in Harvard Data Science Review. I learned a lot by taking it on, including how hard speaking across disciplines is. I'm glad it's finally out! PS Comes w/ a lively discussion ;-) bit.ly/3yeQ2Qr
- It's election forecast time! But can we take @FiveThirtyEight or @Economist forecasts at face value? What are forecasters' goals & incentives, especially post 2016 disillusionment? Excited to share a new paper w/ @StatModeling & @CBWlezien stat.columbia.edu/~gelman/resear… A thread!
- New on Multiple Views: First two posts in @mjskay's and my series on uncertainty visualization We start with what we mean by uncertainty in vis, then walk through techniques from vis research and beyond, discussing trade-offs bit.ly/2Ms5tkB bit.ly/2wCdqZS
- There are 2 things I wish more consumers of experiments grasped: 1. When researchers are incentivized to report significant results, low power reduces the likelihood that a significant result reflects a true effect (it's probably overestimated) shorturl.at/cfNRS 1/
- I'm very happy to be receiving a Microsoft Faculty Fellowship for the work we've been doing on uncertainty vis and belief elicitation for interactive visualization
- New Multiple Views post: @SteveFranconeri's chart chooser formalizes visualization guidelines to make it easy to find the best chart given the structure of your data bit.ly/2Vw3Tla @vis_research
- Occurs to me that one of the most productive (but rare?) ways to increase the standing of women and other marginalized groups in computing is to invite them to collaborate with you, especially if you're in the elite of your field 1/4
- I wrote an op-ed on my fears around COVID-19 uncertainty communication early in lockdown, to help myself figure out what I thought & out of frustration that my first uncertainty op-ed was held hostage by the NYT Happy to see it finally published! blogs.scientificamerican.com/observations/l…
- Like probably many other women in HCI, I’ve found the prominence of Walter Lasecki in my thread very triggering, to the point I've been more irritable & angry in my daily activities than I can remember being in awhile. And I've known about these allegations for months! 1/
- I'm honored to be named Ginni Rometty Assoc Prof @NorthwesternCS for my work on improving data-driven decisions under uncertainty. Much more to do toward understanding (& formalizing) how vis & interaction increase trust & best responding to predictions! bit.ly/3D2FIxb
- I'm seeking a new PhD student at @NorthwesternCS! Interests include modeling visual analysis/learning from data under uncertainty, interface-mediated feedback loops, implications & decision making w/differential privacy. Join my lab w/@mjskay! Pls retweet! mucollective.northwestern.edu
- Excited to have received an NSF CAREER award to rethink the role of prior knowledge in visualization interactionA research project led by the iSchool's @JessicaHullman has earned a $523,000 grant to study and improve the effectiveness of data visualizations: ischool.uw.edu/news/2018/04/j… #Dataviz
