Su Chu

Assistant Prof. @harvardmed and @BrighamWomens. Statistics, epidemiology, integrative 'omics, & causal inference.

[Su Chu @scoutlier]

@marciaixchel Congratulations, Marcia! Well done!

[Su Chu @scoutlier]

I signed up, I hope others do too! Let's all chip in to assist with syndromic surveillance.

[Xihong Lin @XihongLin]

Excited to launch an app HowWeFeel for self-reported COVID-19 symptoms, health and testing information by engaging the community to jointly fight for #COVID19 w/ @zhangf, @8en,@kinggary,@weallen1, @bitdrift and others. Please RT and participate at https://t.co/UOn5SsgkkU

[Xihong Lin @XihongLin]

Analysis results of 25,000+ lab-confirmed COVID-19 cases in Wuhan till Feb 18: epidemiological characteristics and non-pharmaceutical Intervention effects. The key results are summarized below and hope they are useful for the strategies in US and others https://t.co/ajl4MZUubi

[Harvard Biostatistics @HarvardBiostats]

WATCH 3/13/20 webinar: @XihongLin presents the "Analysis of 25, 000+ COVID-19 cases in Wuhan: Epidemiological characteristics and effects of non-pharmaceutical interventions" https://t.co/x1eKgKManc. Updated slides: https://t.co/ZJ95uXpC7j #COVIDー19 #coronavirus @HarvardChanSPH

[Palli Thordarson @PalliThordarson]

1/25 Part 1 - Why does soap work so well on the SARS-CoV-2, the coronavirus and indeed most viruses? Because it is a self-assembled nanoparticle in which the weakest link is the lipid (fatty) bilayer. A two part thread about soap, viruses and supramolecular chemistry #COVID19

[Su Chu @scoutlier]

@tammy_jiang Whoo hoo, go Tammy!

[Su Chu @scoutlier]

@SaraSuliman13 @NextEinsteinFor @DrTolullah Amazing news, congratulations, Sara!

[Su Chu @scoutlier]

Love that #IBC2018 is specifically honoring diff women from diff stats fields from diff countries throughout conference #womeninstatistics

[Craig Hersh MD @craig_hersh]

Contact me if you're inspired to study #COPD genetics after #ATS2018.
Post-doc fellowship available. Please RT. https://t.co/9OBQXbDEli

[Su Chu @scoutlier]

Gonna be famous! @yi_zhao1026 https://t.co/mOEsZr1uBP

[Irene Steves @i_steves]

Loved the cooking metaphors 🔪🎂 in @hadleywickham's @RLadiesSF talk yesterday! but the best part was watching Hadley speed-code live and learning neat tricks like this:

[Rafael Irizarry @rafalab]

Happy #MayFourth #rstats
par(bg=1,fg="white")
x<-0.5->y
z<-"|-o-|"
s<-cbind(runif(50),runif(50))
m<-c(-1,1)/20
while(TRUE){
rafalib::nullplot(xaxt="n",yaxt="n",bty="n")
points(s,pch=".")
text(x,y,z, cex=4)
x<-pmin(pmax(x+sample(m,1),0),1)
y<-pmin(pmax(y+sample(m,1),0),1)
}

[Ken Rothman @ken_rothman]

This is an excellent summary of state of the art methods to control for time-varying confounding, by Mansournia et al, in the BMJ: https://t.co/IpQTSaFX5f

[Nicholas Tierney @nj_tierney]

Want to do the following in #rstats?

- Merge multiple PDF files
- Split single input PDF file into individual pages
- Remove selected pages from a file
- Rename multiple files in a directory

Check out "staplr" (https://t.co/ZKarmZsBQM) by @pridiltal

[Bruce Swihart @bruce_swihart]

COPY file *pathnames* in macOS.

In https://t.co/pmJqEiNYYG:

1. Right-click the item
2. Hold alt ⌥ key down
3. Click Copy "item" as Pathname

Done. Now you have the path, as text, in your clipboard.

Thanks, @theonlygusti! source: https://t.co/VMlbjZ8doL

[Allan Birnbaum @AllanBirnbaum]

My rose has a leaf,
Its thorns are a terror
Please keep your belief–
I want probabilities of error

[Matthew Stephens @mstephens999]

new from my lab, flash for sparse matrix factorization:
preprint: https://t.co/pcjkSJpGGJ
software: https://t.co/byuyeTj6m9
flash can find mixtures of dense and sparse factors, like these tissue-shared and tissue-specific effects in GTEx. Try it out!

[Su Chu @scoutlier]

For when and how to reference causal effects/modeling/inference -- see thread. https://t.co/KeFQtUu48u

[Miguel Hernán @_MiguelHernan]

@f2harrell Glad you asked this question because
- many journals have banned the word "causal"
- many scientists practice a form of self-censorship: they replace the C-word by “association”, even when they are obviously trying to quantify causal effects.

My response & some comments below

[Brian Nosek (@[email protected]) @BrianNosek]

An excellent response to well-intentioned but ineffective "implicit bias training". There is little evidence that such trainings reduce implicit bias, or that, even if they did, that the change in implicit bias addresses the outcomes of interest for diversity and inclusion. https://t.co/cRsbtgtl4L