Advik Shreekumar (@AdvikSh) header
Advik Shreekumar (@AdvikSh) profile picture

PhD student @MITEcon «- @UChiUrbanLabs «- @CivisAnalytics Late PhD potpourri: behavioral + methods. Putting the yo in RStudio.

Boston, MA
Joined June 2009
Photo by Advik Shreekumar (@AdvikSh)Photo by Advik Shreekumar (@AdvikSh)Photo by Advik Shreekumar (@AdvikSh)Photo by Advik Shreekumar (@AdvikSh)Photo by Advik Shreekumar (@AdvikSh)Photo by Advik Shreekumar (@AdvikSh)Photo by Advik Shreekumar (@AdvikSh)Photo by Advik Shreekumar (@AdvikSh)Photo by Advik Shreekumar (@AdvikSh)Photo by Advik Shreekumar (@AdvikSh)
I’m thrilled to announce Conformal Risk Control: a way to bound quantities other than coverage with conformal prediction. arxiv.org/abs/2208.02814 Check out the worked examples in CV and NLP! The best part is: it’s exactly the same algorithm as split conformal prediction🤯🧵1/5
2
29
6
139
Here's a 1-minute video showing--in the words of visualization hero @EdwardTufte--the "compared to what" factor for today's amazing @NASAWebb image. Thanks to the @WWTelescope for making this possible, and @ADavidWeigel. I only wish @NASA and @POTUS had shown the image this way!
271
5,920
2,170
17,125
NEW PREPRINT The increasingly popular class imbalance approaches (such as SMOTE) for risk prediction modeling: they are likely to do more harm than good arxiv.org/abs/2202.09101
14
106
35
452
Advik Shreekumar retweeted
88
2,745
478
17,395
Advik Shreekumar retweeted
Map of the Earth if it was a bear
79
961
103
11,805
Advik Shreekumar retweeted
Something cool in the works: Save your lunch hours (12-1pm ET) Mondays 7/19-8/16 for a research webinar series on how to reduce violent crime. Lots of short talks to get you up to speed on the evidence you should know! Organized w the fine folks at @NiskanenCenter. More soon!
1
17
3
54
I'm excited to share that my paper "Does Meeting Style Matter" has been published in the @apsrjournal. The study leverages a school district that embedded direct response to public comment into its meeting structure. Here's the paper and a long thread: doi.org/10.1017/S00030554210…
25
67
43
316
Advik Shreekumar retweeted
why did they call it 'ethical AI' and not just 'principled component analysis'
11
83
7
820
Advik Shreekumar retweeted
THREAD: On the role of bad social science in the Vietnam War—and how not understanding local conditions, and thinking through endogeneity (yes, endogeneity) can lead to policy conclusions with tragic, deadly consequences.
18
286
100
829
hard to believe this is real: my first ever solo authored paper was just accepted to @apsrjournal 🥲 during the 2019 general elections in India when political misinformation was rife, I designed a field experiment to combat it 🧵1/
55
130
24
1,084
Advik Shreekumar retweeted
My paper on how changing government guidance affects the public’s beliefs — joint with @AdvikSh and @plvautrey — is now up on the @JPubEcon website! This is my first publication (!), and I am excited to share a little about it (1/N) sciencedirect.com/science/ar…
1
16
3
83
Where frequentism really excels is at easy semiparametrics: the intersection of "I don't know" and "I don't care." You can focus on the things you need to get right like treatment effects and be robust to nuisances like error forms or serial correlation 🧵 teapawt.net/jmwooldridge/status/13…
Replying to @jmwooldridge
But those estimators can be studied from a frequentist perspective, and no strong assumptions are needed. My hesitation with Bayesian methods -- when they differ from classical ones -- is that they are not "robust" in the econometrics sense.
2
10
5
62
Why do they call it a dog park and not a zoom meeting
1
2
10
Advik Shreekumar retweeted
Looking for help: Several profs have contacted me saying they'd like to open their online classes to guest students. So maybe this has legs? I've registered remotestudentexchange.org! Anyone good with MySQL/PHP/whatever the cool kids use to help me build it? No money, just glory.
This Post is from an account that no longer exists. Learn more
7
40
2
130
So some very smart folks have asked about how we would apply the AMIP metric to studies of rare events. This kicked off a discussion of what robustness checks are really for, and I want to take that set of questions seriously in this thread.
6
34
29
202
What's it like to work as an economist at a non-academic research institute, like RAND? This is a question I get a lot, especially during job market season. #EconTwitter I have a new report out today that I think offers one perspective. (thread) 1/ rand.org/pubs/research_repor…
1
14
3
85
Advik Shreekumar retweeted
🚨 New working paper (w/ @aecoppock and #GregHuber) "The Generalizability of Online Experiments Conducted During The COVID-19 Pandemic" 🚨 Now online at @socarxiv (link: https://teapawt.net/t.co/pHSLAEkg54) Short summary 👇
1
47
13
150
Advik Shreekumar retweeted
I couldn’t just walk past this Tweet, so here is some fun #dataviz Scented candles: An unexpected victim of the COVID-19 pandemic 1/n
This Post is from an account that no longer exists. Learn more
250
10,096
3,775
27,353
1/n Some thoughts from a first-year econ phd on how the year is going: It's not going well. Learning over zoom sucks, not being able to have regular and efficient study groups sucks, and there's not much that can change there.
3
7
3
118