Six former Census Bureau directors have called for the agency to withdraw the proposed residency requirements regulations.
thecensusproject.org/wp-content/u...
thecensusproject.org/wp-content/u...
thecensusproject.org
October 29, 2026 at 7:46 PM
October 29, 2026 at 9:00 PM
October 28, 2026 at 12:23 AM
#roshtein #kickban
statsky 🔨 BANNED by MUTTUUUU (after 4s)
"for that: discord .gg/k-bot"
kick.com/roshtein
statsky 🔨 BANNED by MUTTUUUU (after 4s)
"for that: discord .gg/k-bot"
kick.com/roshtein
October 31, 2026 at 11:51 PM
#roshtein #kickban
statsky 🔨 BANNED by MUTTUUUU (after 4s)
"for that: discord .gg/k-bot"
kick.com/roshtein
statsky 🔨 BANNED by MUTTUUUU (after 4s)
"for that: discord .gg/k-bot"
kick.com/roshtein
#roshtein #kickban
statsky 🔨 BANNED by MUTTUUUU (after 4s)
"for that: discord .gg/k-bot"
kick.com/roshtein
statsky 🔨 BANNED by MUTTUUUU (after 4s)
"for that: discord .gg/k-bot"
kick.com/roshtein
October 31, 2026 at 11:51 PM
#roshtein #kickban
statsky 🔨 BANNED by MUTTUUUU (after 4s)
"for that: discord .gg/k-bot"
kick.com/roshtein
statsky 🔨 BANNED by MUTTUUUU (after 4s)
"for that: discord .gg/k-bot"
kick.com/roshtein
Come work with me!
The Dept. of Math & Stats at the University of Alaska Fairbanks is hiring TWO tenure-track tripartite(teaching/research/service) statisticians!
We're a great department! and you will have the opportunity to reshape the stats program.
#statsky
careers.alaska.edu/jobs/assista...
The Dept. of Math & Stats at the University of Alaska Fairbanks is hiring TWO tenure-track tripartite(teaching/research/service) statisticians!
We're a great department! and you will have the opportunity to reshape the stats program.
#statsky
careers.alaska.edu/jobs/assista...
Assistant Professor of Statistics - Fairbanks, Alaska, United States
The Department of Mathematics and Statistics at University of Alaska Fairbanks seeks two tenure-track Assistant Professors of Statistics to begin Fall 2027. Successful candidates in Statistics and Dat...
careers.alaska.edu
October 28, 2026 at 5:35 PM
Come work with me!
The Dept. of Math & Stats at the University of Alaska Fairbanks is hiring TWO tenure-track tripartite(teaching/research/service) statisticians!
We're a great department! and you will have the opportunity to reshape the stats program.
#statsky
careers.alaska.edu/jobs/assista...
The Dept. of Math & Stats at the University of Alaska Fairbanks is hiring TWO tenure-track tripartite(teaching/research/service) statisticians!
We're a great department! and you will have the opportunity to reshape the stats program.
#statsky
careers.alaska.edu/jobs/assista...
The more I read "Factfulness" by Hans Rosling, the more I realize that part of modern society's issues would be solved by teaching that book at school.
#statsky
#statsky
October 27, 2026 at 7:19 AM
The more I read "Factfulness" by Hans Rosling, the more I realize that part of modern society's issues would be solved by teaching that book at school.
#statsky
#statsky
🎉 New preprint on Estimating Stability of Individual-Specific Predictions using Bootstrap and CV Resampling
One important takeaway: don’t use standard non-parametric bootstrap as your default for all model types! It performed poorly for our Random Forest models.
arxiv.org/abs/2610.33744
#Statsky
One important takeaway: don’t use standard non-parametric bootstrap as your default for all model types! It performed poorly for our Random Forest models.
arxiv.org/abs/2610.33744
#Statsky
October 27, 2026 at 10:00 AM
🎉 New preprint on Estimating Stability of Individual-Specific Predictions using Bootstrap and CV Resampling
One important takeaway: don’t use standard non-parametric bootstrap as your default for all model types! It performed poorly for our Random Forest models.
arxiv.org/abs/2610.33744
#Statsky
One important takeaway: don’t use standard non-parametric bootstrap as your default for all model types! It performed poorly for our Random Forest models.
arxiv.org/abs/2610.33744
#Statsky
How the log-rank test works, read from survdiff output on 228 people, and why crossing survival curves cost it power: celeus.ai/blog/log-ran... #statsky #survivalanalysis
Log-rank test: comparing survival curves in R - Celeus
How the log-rank test compares survival curves, how to read survdiff output, why crossing curves cost it power, and when to use a Peto-Peto weighting.
celeus.ai
October 30, 2026 at 9:13 PM
How the log-rank test works, read from survdiff output on 228 people, and why crossing survival curves cost it power: celeus.ai/blog/log-ran... #statsky #survivalanalysis
A/B Tests, Double Machine Learning, Synthetic Control
What to use, when, and why, under real-world constraints?
Later today, at ODSC AI West, I'm running a session on causal impact evaluation, based on a project I recently worked on for a Fortune 500 client.
1/
#CausalSky #StatSky #MLSky
What to use, when, and why, under real-world constraints?
Later today, at ODSC AI West, I'm running a session on causal impact evaluation, based on a project I recently worked on for a Fortune 500 client.
1/
#CausalSky #StatSky #MLSky
October 29, 2026 at 10:28 AM
A/B Tests, Double Machine Learning, Synthetic Control
What to use, when, and why, under real-world constraints?
Later today, at ODSC AI West, I'm running a session on causal impact evaluation, based on a project I recently worked on for a Fortune 500 client.
1/
#CausalSky #StatSky #MLSky
What to use, when, and why, under real-world constraints?
Later today, at ODSC AI West, I'm running a session on causal impact evaluation, based on a project I recently worked on for a Fortune 500 client.
1/
#CausalSky #StatSky #MLSky
Dropping censored observations, or counting the cut-off as a death, both bias survival estimates. On 26 women from a public trial, 14 were censored and held the ten longest follow-up times: celeus.ai/blog/censore...
#statsky #survivalanalysis
#statsky #survivalanalysis
Censored data: right, left and interval censoring - Celeus
What censored data are, how right, left and interval censoring differ, and why dropping censored observations biases survival estimates, on public trial data.
celeus.ai
October 30, 2026 at 1:23 AM
Dropping censored observations, or counting the cut-off as a death, both bias survival estimates. On 26 women from a public trial, 14 were censored and held the ten longest follow-up times: celeus.ai/blog/censore...
#statsky #survivalanalysis
#statsky #survivalanalysis
Hey folks,
On Thursday I'm running a workshop at Open Data Science Conference (ODSC).
We'll work through a real-world-inspired example of estimating the effect of a sales-rep-facing recommender system on revenue.
1/
#CausalSky #StatSky #EconSky #MLSky
On Thursday I'm running a workshop at Open Data Science Conference (ODSC).
We'll work through a real-world-inspired example of estimating the effect of a sales-rep-facing recommender system on revenue.
1/
#CausalSky #StatSky #EconSky #MLSky
October 26, 2026 at 9:25 AM
Hey folks,
On Thursday I'm running a workshop at Open Data Science Conference (ODSC).
We'll work through a real-world-inspired example of estimating the effect of a sales-rep-facing recommender system on revenue.
1/
#CausalSky #StatSky #EconSky #MLSky
On Thursday I'm running a workshop at Open Data Science Conference (ODSC).
We'll work through a real-world-inspired example of estimating the effect of a sales-rep-facing recommender system on revenue.
1/
#CausalSky #StatSky #EconSky #MLSky
Agents can fail miserably with counterfactual reasoning tasks.
How good are we, humans, at them?
This week in Causal Python Weekly:
1/
#CausalSky #StatSky #EconSky #EpiSky #MLSky
How good are we, humans, at them?
This week in Causal Python Weekly:
1/
#CausalSky #StatSky #EconSky #EpiSky #MLSky
October 25, 2026 at 9:58 AM
Agents can fail miserably with counterfactual reasoning tasks.
How good are we, humans, at them?
This week in Causal Python Weekly:
1/
#CausalSky #StatSky #EconSky #EpiSky #MLSky
How good are we, humans, at them?
This week in Causal Python Weekly:
1/
#CausalSky #StatSky #EconSky #EpiSky #MLSky
Big payouts come with small chances. On a European wheel, one number pays 35 to 1 but wins only 2.7% of the time. Red or black pays 1 to 1 and wins 48.6% of the time. The main bets in one chart: therouletteassistant.com/roulette/payouts/ #probability #statsky
October 26, 2026 at 8:32 AM
Big payouts come with small chances. On a European wheel, one number pays 35 to 1 but wins only 2.7% of the time. Red or black pays 1 to 1 and wins 48.6% of the time. The main bets in one chart: therouletteassistant.com/roulette/payouts/ #probability #statsky
#adinross #kickban
statsky 🔨 BANNED by shids (after 2m05s)
"sweet rush bonanza-mmafighter010"
kick.com/adinross
statsky 🔨 BANNED by shids (after 2m05s)
"sweet rush bonanza-mmafighter010"
kick.com/adinross
October 22, 2026 at 9:19 AM
#adinross #kickban
statsky 🔨 BANNED by shids (after 2m05s)
"sweet rush bonanza-mmafighter010"
kick.com/adinross
statsky 🔨 BANNED by shids (after 2m05s)
"sweet rush bonanza-mmafighter010"
kick.com/adinross
October 21, 2026 at 3:42 PM
Fare evasion changes the answer. On paid rides alone, subway grand larceny looks level with 2019 and robbery higher. Count every entry, paid or not, and grand larceny is 13% lower, perhaps by chance, and robbery about the same.
How we measured it: sparkydata.dev/articles/is-the-… 📊 #statsky
How we measured it: sparkydata.dev/articles/is-the-… 📊 #statsky
October 23, 2026 at 1:03 PM
Fare evasion changes the answer. On paid rides alone, subway grand larceny looks level with 2019 and robbery higher. Count every entry, paid or not, and grand larceny is 13% lower, perhaps by chance, and robbery about the same.
How we measured it: sparkydata.dev/articles/is-the-… 📊 #statsky
How we measured it: sparkydata.dev/articles/is-the-… 📊 #statsky
Is it possible, for each study, to estimate uncertainty around where the inflection point is, then to take these point estimates with their associated confidence intervals, and test for heterogeneity, as you would in a meta analysis?
#statsky
#statsky
Statistical friends, imagine you had data on the association betwen two continuous variables across a number of different settings
Slope levels off at some point
What would the best approach to testing whether the inflection point differed between settings?
#statsky
Slope levels off at some point
What would the best approach to testing whether the inflection point differed between settings?
#statsky
October 19, 2026 at 3:24 PM
Is it possible, for each study, to estimate uncertainty around where the inflection point is, then to take these point estimates with their associated confidence intervals, and test for heterogeneity, as you would in a meta analysis?
#statsky
#statsky
October 18, 2026 at 12:21 PM
Pedal-bike deaths in New York fell from 28 in 2019 to 7 in 2025. Before anyone claims a win: e-bikes got their own police label only in August 2019, so some riders may be filed elsewhere. Bicycles and e-devices together: still up.
How we checked: sparkydata.dev/articles/who-sti… 📊 #statsky
How we checked: sparkydata.dev/articles/who-sti… 📊 #statsky
October 20, 2026 at 10:00 PM
Pedal-bike deaths in New York fell from 28 in 2019 to 7 in 2025. Before anyone claims a win: e-bikes got their own police label only in August 2019, so some riders may be filed elsewhere. Bicycles and e-devices together: still up.
How we checked: sparkydata.dev/articles/who-sti… 📊 #statsky
How we checked: sparkydata.dev/articles/who-sti… 📊 #statsky
Reading Breiman again and wondering whether this dichotomy is still valid. His take homes are excellent: 1) focus on the problem and the data; 2) if you are after the DGP, predictive accuracy matters most. But the point of much applied work is to propose and test a mechanism. (1/4) #statsky #episky
Subtweeting yesterday’s AI discourse, I wrote about Breiman’s Two Cultures paper and what we want from our forecasting models.
Predictive Models Are Never Clean
Revisiting Breiman’s Two Cultures paper and what we want from our forecasting models.
www.argmin.net
October 17, 2026 at 3:11 AM
In the famous 1913 run, black came up 26 times in a row. One player says red is due. Another says black is hot. On a fair wheel, both are wrong. Spin 27 is like spin 1: red 18 in 37, black 18 in 37, zero 1 in 37. therouletteassistant.com/roulette/gamblers-fallacy/ #statistics #statsky
The gambler's fallacy at the roulette wheel: no memory, no due numbers
Why streaks feel meaningful and are not: the independence math, real streak lengths in 100,000 simulated sessions, and a safe way to test it.
therouletteassistant.com
October 18, 2026 at 12:32 PM
In the famous 1913 run, black came up 26 times in a row. One player says red is due. Another says black is hot. On a fair wheel, both are wrong. Spin 27 is like spin 1: red 18 in 37, black 18 in 37, zero 1 in 37. therouletteassistant.com/roulette/gamblers-fallacy/ #statistics #statsky
Ten betting systems, 100,000 simulated European roulette sessions each. Every one landed on the same line: about 1 point lost for every 37 points bet. That line is the house edge. therouletteassistant.com/lab-notes/one-million-sessions/ #dataviz #statsky
October 16, 2026 at 5:36 PM
Ten betting systems, 100,000 simulated European roulette sessions each. Every one landed on the same line: about 1 point lost for every 37 points bet. That line is the house edge. therouletteassistant.com/lab-notes/one-million-sessions/ #dataviz #statsky
Latent class analysis is astrology for scientists
#statsky
#statsky
Mercury is in Retrograde Reaction
ALT: Mercury is in Retrograde Reaction
static.klipy.com
October 15, 2026 at 7:42 AM
Latent class analysis is astrology for scientists
#statsky
#statsky