Mark Rieke
@markjrieke.bsky.social
certified big nerd | big #rstats dweeb | accidental python fan | he/him
thedatadiary.net
thedatadiary.net
Pinned
Mark Rieke
@markjrieke.bsky.social
· May 25
Here are some of the photos from hiking rim to rim. It’s difficult to capture in a photo (or words) the sense of wonder and majesty imbued by this park. The scale is humbling
sometimes they put the yee haw right into the programming language
February 15, 2026 at 10:20 PM
sometimes they put the yee haw right into the programming language
olympic bobsled should add podracing noises send tweet
February 15, 2026 at 5:21 PM
olympic bobsled should add podracing noises send tweet
Reposted by Mark Rieke
Also, Yglesias lecturing people on uncertainty when it comes to interpreting polling data is laughable
February 15, 2026 at 5:11 PM
Also, Yglesias lecturing people on uncertainty when it comes to interpreting polling data is laughable
Reposted by Mark Rieke
Or, as written up in SIGBOIVK last year:
February 15, 2026 at 3:42 PM
Or, as written up in SIGBOIVK last year:
Reposted by Mark Rieke
The stupidity, the criminal vandalism, the wanton destruction of information involved in dichotomisation
statmodeling.stat.columbia.edu/2026/02/14/t...
statmodeling.stat.columbia.edu/2026/02/14/t...
The stupidity, the criminal vandalism, the wanton destruction of information involved in dichotomisation | Statistical Modeling, Causal Inference, and Social Science
statmodeling.stat.columbia.edu
February 14, 2026 at 6:03 PM
The stupidity, the criminal vandalism, the wanton destruction of information involved in dichotomisation
statmodeling.stat.columbia.edu/2026/02/14/t...
statmodeling.stat.columbia.edu/2026/02/14/t...
after my sufficient gamma crashout last night, I've done some simulations & realized:
- the pdf below works for a sufficient gamma
- the original pdf I shared ALSO works for a sufficient statistic under a different parameterization
- weve already done this at work under a DIFFERENT parameterization
- the pdf below works for a sufficient gamma
- the original pdf I shared ALSO works for a sufficient statistic under a different parameterization
- weve already done this at work under a DIFFERENT parameterization
if I just sit down and bake out what a repeated product of the gamma pdf is I end up with this --- needs to be checked with simulation, but certainly passes the gut check
February 13, 2026 at 5:28 PM
after my sufficient gamma crashout last night, I've done some simulations & realized:
- the pdf below works for a sufficient gamma
- the original pdf I shared ALSO works for a sufficient statistic under a different parameterization
- weve already done this at work under a DIFFERENT parameterization
- the pdf below works for a sufficient gamma
- the original pdf I shared ALSO works for a sufficient statistic under a different parameterization
- weve already done this at work under a DIFFERENT parameterization
if I just sit down and bake out what a repeated product of the gamma pdf is I end up with this --- needs to be checked with simulation, but certainly passes the gut check
February 13, 2026 at 4:14 AM
if I just sit down and bake out what a repeated product of the gamma pdf is I end up with this --- needs to be checked with simulation, but certainly passes the gut check
someone also has pointed out that x ~ Gamma(alpha, theta) is equivalent to sum(x) ~ Gamma(n * alpha, theta). At one point I had convinced myself that there was a loss of individual information here, but now I can't quite see why I came to that conclusion --- gotta simulate some stuff to see !
February 12, 2026 at 11:36 PM
someone also has pointed out that x ~ Gamma(alpha, theta) is equivalent to sum(x) ~ Gamma(n * alpha, theta). At one point I had convinced myself that there was a loss of individual information here, but now I can't quite see why I came to that conclusion --- gotta simulate some stuff to see !
spitballing a function and maybe this works? would need to test it out against a non-sufficient gamma
February 12, 2026 at 10:28 PM
spitballing a function and maybe this works? would need to test it out against a non-sufficient gamma
it is a plot, we will get all of you, eventually, to embrace bayesian inference, resistance is futile, etc. etc.
(this is also part of my secret plan to get people to be bayesian)
February 12, 2026 at 1:12 PM
it is a plot, we will get all of you, eventually, to embrace bayesian inference, resistance is futile, etc. etc.
Reposted by Mark Rieke
New gaussian process slides going well
February 11, 2026 at 4:48 PM
New gaussian process slides going well
I have reached the "writing yaml specifications" stage of my career
February 10, 2026 at 5:14 PM
I have reached the "writing yaml specifications" stage of my career
why is jazz so fuckin' good
February 8, 2026 at 5:39 PM
why is jazz so fuckin' good
Reposted by Mark Rieke
"Is my sample large enough for Bayesian statistics?" is a weird question coming from someone who only ever uses frequentist stats that are only approximately correct for large samples.
February 8, 2026 at 3:26 PM
"Is my sample large enough for Bayesian statistics?" is a weird question coming from someone who only ever uses frequentist stats that are only approximately correct for large samples.
"production notebook" is an oxymoron and should be shunned from pleasant society
February 6, 2026 at 11:28 PM
"production notebook" is an oxymoron and should be shunned from pleasant society
this series is mostly horny nonsense (positive review), but this scene is legitimately incredible and moving
February 6, 2026 at 5:25 PM
this series is mostly horny nonsense (positive review), but this scene is legitimately incredible and moving
nothing so captivating as watching someone exert complete mastery over a skill with ease
Ian McKellen performs “The Strangers’ Case” speech from “Sir Thomas More” on Colbert.
February 6, 2026 at 1:00 PM
nothing so captivating as watching someone exert complete mastery over a skill with ease
I've been learning a new programming paradigm and even though I'm very bad and slow it's incredibly fun because the dopamine hit of figuring out something simple for the first time simply cannot be matched
February 5, 2026 at 3:12 AM
I've been learning a new programming paradigm and even though I'm very bad and slow it's incredibly fun because the dopamine hit of figuring out something simple for the first time simply cannot be matched
there's a certain genre of guy (and it is mostly men) who's good at X technical thing but assumes that means they're good at ALL technical things
"data grifter" is a good term for this when it's someone who's good at building ML models but confidently makes pretty basic causal inference mistakes
"data grifter" is a good term for this when it's someone who's good at building ML models but confidently makes pretty basic causal inference mistakes
I hattttttttteeeeeee to tell you how absolutely prevalent this form of data grifting is. We are in an era of data grifting around how people talk about interventions specifically
basically yglesias, jain, jentleson, WelcomePAC donors etc all base their prescriptions for Dems on a statistical model that nobody can replicate, and refuse to acknowledge (a) that uncertainty dwarfs the detectable effect of moderation & (b) that the last year has proved their views wrong!!
February 4, 2026 at 2:31 PM
there's a certain genre of guy (and it is mostly men) who's good at X technical thing but assumes that means they're good at ALL technical things
"data grifter" is a good term for this when it's someone who's good at building ML models but confidently makes pretty basic causal inference mistakes
"data grifter" is a good term for this when it's someone who's good at building ML models but confidently makes pretty basic causal inference mistakes
Reposted by Mark Rieke
I hattttttttteeeeeee to tell you how absolutely prevalent this form of data grifting is. We are in an era of data grifting around how people talk about interventions specifically
basically yglesias, jain, jentleson, WelcomePAC donors etc all base their prescriptions for Dems on a statistical model that nobody can replicate, and refuse to acknowledge (a) that uncertainty dwarfs the detectable effect of moderation & (b) that the last year has proved their views wrong!!
February 4, 2026 at 2:05 AM
I hattttttttteeeeeee to tell you how absolutely prevalent this form of data grifting is. We are in an era of data grifting around how people talk about interventions specifically
Reposted by Mark Rieke
It is frustrating that half of the folks debating moderation are backing up their claims with open source statistical analyses whereas the other half gets away with, what appears to be, basic statistical mistakes with closed-source methods
February 3, 2026 at 6:10 PM
It is frustrating that half of the folks debating moderation are backing up their claims with open source statistical analyses whereas the other half gets away with, what appears to be, basic statistical mistakes with closed-source methods
Reposted by Mark Rieke
If 1/10 of the money and effort invested in investigating and debating moderation effects was invested in some of the other electoral strategies Bonica and Grumbach discussed (I'd emphasize corruption, but am open to many alternative ideas), the Democratic party would already be better off
February 3, 2026 at 5:59 PM
If 1/10 of the money and effort invested in investigating and debating moderation effects was invested in some of the other electoral strategies Bonica and Grumbach discussed (I'd emphasize corruption, but am open to many alternative ideas), the Democratic party would already be better off
Reposted by Mark Rieke
anyhoo, it seems that we're doomed to run through this cycle once every few months until the end of time
February 3, 2026 at 6:19 PM
anyhoo, it seems that we're doomed to run through this cycle once every few months until the end of time
Reposted by Mark Rieke
some additional reading
www.gelliottmorris.com/p/data-over-...
www.gelliottmorris.com/p/the-cost-o...
www.gelliottmorris.com/p/more-evide...
www.bostonreview.net/forum/how-no...
www.gelliottmorris.com/p/data-over-...
www.gelliottmorris.com/p/the-cost-o...
www.gelliottmorris.com/p/more-evide...
www.bostonreview.net/forum/how-no...
Data over dogma: A reply to Matt Yglesias
Further thoughts and data on the declining electoral benefit to ideological moderation.
www.gelliottmorris.com
February 3, 2026 at 6:12 PM