Dave Childers

Stats & Data Science. Wordles and parenting with @pedpallimama

[Dave Childers @dmchilders]

@GantLaborde x <- 1:10
x[which(x) < 4]

[Dave Childers @dmchilders]

@drob for the first time ever, answer is "yes to both"

[Dave Childers @dmchilders]

@narwhaldevil95 @levantinenative @rabbisandra What a vile statement. My family attended zoom services by Rabbi Sandra, and they were wonderful. She's doing very important work. If you'd like to join me in supporting her meaningful work, I believe you can find her patreon here: https://t.co/TKm6atNhnS

[Dave Childers @dmchilders]

@JoeEmison My personal experience:

Number of data scientists who use "AI" to describe what they do: plenty
Number of effective data scientists who use "AI" to describe what they do: zero

[Dave Childers @dmchilders]

@JoeEmison In large orgs, people can make a career out of moving into a trendy role (e.g. DS), dazzling others with hype (AI), not deliver any value, and then rinsing and repeating every couple of years.

[Dave Childers @dmchilders]

@pedpallimama "do you" or "should you"?

[Julie SW Childers, MD PhD FAAP @pedpallimama]

Wordle 206 2/6

🟨🟩⬜⬜⬜
🟩🟩🟩🟩🟩

Team effort. On the couch. ⁦@dmchilders⁩

[Dave Childers @dmchilders]

@eamcvey "welcome to the 2021-22 academic school year where we will prepare your children for success in a digitized world. now, please write backup contact information on these 10 separate forms for the classroom, school office, district office, ..."

[John Burn-Murdoch @jburnmurdoch]

Here are the latest case trajectories:

• Japan’s flatter curve now felt to be due to effective isolation of elderly, more than lower testing
• Underscores trend of Asian countries responding effectively

Keep up to date with latest developments here: https://t.co/VcSZISFxzF

[Dave Childers @dmchilders]

@KeckMedUSC Are you following the advice of the surgeon general and cancelling elective procedures: https://t.co/NEBvxwfFLw

[U.S. Surgeon General @Surgeon_General]

Hospital & healthcare systems, PLEASE CONSIDER STOPPING ELECTIVE PROCEDURES until we can #FlattenTheCurve! 👇🏽

Each elective surgery you do:

1) Brings possible #Coronavirus to your facilities
2) Pulls from PPE stores
3) Taxes personnel who may be needed for #COVIDー19 response

[Dave Childers @dmchilders]

@AmerAcadPedDent The @AmerDentalAssn just made a recommendation to postpone all elective dental procedures *nationwide*. Kids are spreaders. Pls do your part to #FlattenTheCurve of the #COVID19 pandemic by recommending the same: https://t.co/OFykRP4rHw

[ADA @AmerDentalAssn]

The ADA recognizes the unprecedented and extraordinary circumstances dentists face related to #COVID19. The ADA recommends dentists nationwide postpone elective procedures for the next three weeks. Please visit https://t.co/VmvCLo4oS9 for full statement and latest information.

[Dave Childers @dmchilders]

@FIGAROCROWNS @AmerAcadPedDent @AmerAcadPedDent - it's great that the program is cancelled. But can you make the right call and declare that dental practices should be closed except for emergency care due to #COVID19 . Dental offices are very high risk and kids are spreaders: https://t.co/aHxIYPAKDA

[Atul Gawande @Atul_Gawande]

We are ringing the alarm bell, folks. The US is 10-14 days behind Italy. We're not doing enough to change that trajectory.

They had 368 deaths yesterday. For a nation of 60 million, it was enough to raise their all-cause mortality rate 20%.

We can avoid that.

#StayHome

[Dave Childers @dmchilders]

@SAISnews The Lerner School used these videos at our board orientation. Incredibly useful resources.

[Dave Childers @dmchilders]

@f2harrell Question: Do you believe p-values are ever “just fine”? What about a test for covariate balance in observational studies? P-value = 0.25 -> this observational study has more balance than 25% of randomized studies. In that case, the null may be meaningful.

[Dave Childers @dmchilders]

@taraskaduk I would be curious to see what the responses would be if you tagged #Python instead of #rstats -- my guess is a lot more votes for CS

[Dave Childers @dmchilders]

Base #rstats has lots of type unstable functions, but `diag` is especially peculiar: give diag() a vector and get a matrix, give it a matrix and get a vector. Any other functions have a similarly inverted input/output relationship?

[Dave Childers @dmchilders]

@f2harrell How would you classify regression/smooth splines? To me, those feel like statistical models -- they are a flexible extension of (G)LMs. But the individual parameters are of less individual interest when working with splines (it's more about the parameters they work together).

[Dave Childers @dmchilders]

@f2harrell @AndrewPiFactori @RSS_DSS I agree that logistic regression is a SM -- but some popular online ML courses unfortunately introduce regression (linear and logistic) as very simple machine learning algorithms.