Adam Bartonicek

PhD student on hiatus. Well-being, stats, & data science. Insufferable R partisan 💻 Monte Carlo maniac 🎲 Always keen for a surf 🏄

[Adam Bartonicek @BartonicekAdam]

@PhDemetri Have you checked out predictive projection feature selection: https://t.co/7Zn1LXd279? Imho it's a really cool & useful Bayesian variable selection method that gets nowhere near the attention it deserves.

[Adam Bartonicek @BartonicekAdam]

@betanalpha

[Adam Bartonicek @BartonicekAdam]

@kdpsinghlab Dang, this is just the perfect amount of cheesy for a stats joke, Karandeep! 😁

[Adam Bartonicek @BartonicekAdam]

@betanalpha Just buy it on credit & tell them you'll pay it of asymptotically

[Alexandre Andorra @alex_andorra]

4th week of 10 tips to improve your #RegressionModels:
"Interpret #Regression Coefficients as Comparisons"
Discover why in ep. 20 of LBS: https://t.co/y9BQ6fBt3Y

Regression and Other Stories @CambridgeUP, by @StatModeling, J. Hill & @avehtari
Next tip next week ;)

[Adam Bartonicek @BartonicekAdam]

@ChelseaParlett (notice that the former spends a lot more time talking about the potato, whereas the latter focuses on the actions)

[Adam Bartonicek @BartonicekAdam]

@ChelseaParlett So lapply/map just seems like a slightly more abstract way of solving the problem.

Instead of:
"For every potato in a bucket, take the potato, peel the potato, and put the potato in a pot.

This:
"Take bucket of potatoes, peel them, and put them in a pot.

[Adam Bartonicek @BartonicekAdam]

@kareem_carr I think that when you talk to another person, you manifest the conviction that there's something in your two sets of experiences that overlaps. How could you ever communicate with someone that does not have even the tiniest bit of experience that matches yours?

[Adam Bartonicek @BartonicekAdam]

@jwiechers @RegaMyEgga @knighton_bob @kareem_carr Searle makes this argument /w computers: computers don't really "compute" anything, they're electrical circuits that take in input & spit out output. Anything can be a computer: a falling rock "computes" t = sqrt(2d/g). However, that doesn't mean computers aren't useful.

[Adam Bartonicek @BartonicekAdam]

@jwiechers @RegaMyEgga @knighton_bob @kareem_carr In that case, if social constructs can be useful, then the fact "x is a social construct" alone doesn't justify disposing of x, right?

[Richard McElreath 🐈‍⬛ @rlmcelreath]

Statistical Rethinking 2nd edition page now lists code conversions for:
* raw Stan+tidyverse
* brms+tidyverse
* PyMC3
* Tensorflow Probability
* Julia & Turing
I know other conversions in the works. If I have missed something, please let me know. https://t.co/jyp2mxBKgC

[Adam Bartonicek @BartonicekAdam]

@Katelinsaurus @betanalpha In my native language (Czech), we have words without wovels, & you'd pronounce "sqrt" as "squirt".

Thanks for coming to my TED talk.

[Prof. Katelin Schutz @Katelinsaurus]

whenever I type `sqrt` while coding, in my mind I always read it as "skrrt" in Cardi B's voice, am I normal

[Adam Bartonicek @BartonicekAdam]

@jamesheathers I'm no evolutionary biologist, but isn't the "genes get selected to increase fitness of the species as a whole" idea debunked?

There must be another reason. My bet is that it's supposed to counter the excessive sweating that arises as a result of forehead space expanding.

[Adam Bartonicek @BartonicekAdam]

@betanalpha F

[Adam Bartonicek @BartonicekAdam]

@kareem_carr Varrus: The probability of my legions not crossing Teutoburg is extremely small
Arminius: Imma update this man's posterior

[Adam Bartonicek @BartonicekAdam]

@nicoleradziwill Because as every programmer knows, more lines = better code

[Adam Bartonicek @BartonicekAdam]

I've got a joke about the mean of the posterior distributions but not sure if it will line up with your expectations

[Adam Bartonicek @BartonicekAdam]

@f2harrell Also, what do you think about precision-based stopping rules? I.e. collecting data until you reach a desired precision/width of credible interval around params of interest? Came across this idea in Kruschke, seems really useful, don't know why it's not used more often.