,
18 tweets,
13 min read
Read on Twitter
1/n
Someone asked me recently what resources I’d recommend for furthering one’s introduction to Bayesian statistics after going through @rlmcelreath's text (xcelab.net/rm/statistical…) and my accompanying project (bookdown.org/connect/#/apps…).
Here are my thoughts:
Someone asked me recently what resources I’d recommend for furthering one’s introduction to Bayesian statistics after going through @rlmcelreath's text (xcelab.net/rm/statistical…) and my accompanying project (bookdown.org/connect/#/apps…).
Here are my thoughts:
2/n
It probably goes without saying, but just in case you missed it, make sure you check out McElreath’s lectures on his text, too youtube.com/channel/UCNJK6…. He has three semester’s worth and they’re overall really great.
It probably goes without saying, but just in case you missed it, make sure you check out McElreath’s lectures on his text, too youtube.com/channel/UCNJK6…. He has three semester’s worth and they’re overall really great.
3/n
And plus, I also like his sand-alone lecture on “Bayesian Statistics without Frequentist Language” . It’s more conceptual than applied, but we could all probably do with a little more philosophy of statistics in our lives.
And plus, I also like his sand-alone lecture on “Bayesian Statistics without Frequentist Language” . It’s more conceptual than applied, but we could all probably do with a little more philosophy of statistics in our lives.
4/n
Recently, I’ve been slowly going through Krushke’s intro text sites.google.com/site/doingbaye…. Compared to McElreath, Kruschke's a touch heavier on the math and his book is organized quite differently. He also covers some great additional topics, such as Bayesian power analyses.
Recently, I’ve been slowly going through Krushke’s intro text sites.google.com/site/doingbaye…. Compared to McElreath, Kruschke's a touch heavier on the math and his book is organized quite differently. He also covers some great additional topics, such as Bayesian power analyses.
5/n
But heads up: Kruschke’s code is a little dated and much heavier on JAGS than Stan—though he does cover Stan a bit. I’m slowly going through the text and converting it to a #brms and #tidyverse format. It’ll be a while before that project is done github.com/ASKurz/Doing-B….
But heads up: Kruschke’s code is a little dated and much heavier on JAGS than Stan—though he does cover Stan a bit. I’m slowly going through the text and converting it to a #brms and #tidyverse format. It’ll be a while before that project is done github.com/ASKurz/Doing-B….
6/n
Still on the applied side, many of us have been waiting eagerly for the revision of Gelman and Hill’s classic text stat.columbia.edu/~gelman/arm/. The original is great in a pinch, but the code is really quite dated and Gelman’s thoughts on things like priors have since changed.
Still on the applied side, many of us have been waiting eagerly for the revision of Gelman and Hill’s classic text stat.columbia.edu/~gelman/arm/. The original is great in a pinch, but the code is really quite dated and Gelman’s thoughts on things like priors have since changed.
7/n
My understanding is that the revision will be split into two volumes, with the first focusing on single-level models and the second focusing on multilevel models stat.columbia.edu/~gelman/regres…. For a preview of the content, check out this online index avehtari.github.io/RAOS-Examples/.
My understanding is that the revision will be split into two volumes, with the first focusing on single-level models and the second focusing on multilevel models stat.columbia.edu/~gelman/regres…. For a preview of the content, check out this online index avehtari.github.io/RAOS-Examples/.
8/n
Relatedly, there is always the authoritative BDA stat.columbia.edu/~gelman/book/. Among the texts I’ve mentioned, this is the most technical and probably most appropriate for budding statisticians. But do check Aki Vehtari’s GitHub repo on code for the text github.com/avehtari/BDA_R…
Relatedly, there is always the authoritative BDA stat.columbia.edu/~gelman/book/. Among the texts I’ve mentioned, this is the most technical and probably most appropriate for budding statisticians. But do check Aki Vehtari’s GitHub repo on code for the text github.com/avehtari/BDA_R…
9/n
If you have a background in SEM, the Mplus team’s lectures on Bayesian SEM are quite nice (i.e., Topic 9 statmodel.com/course_materia…). Though they only show Mplus code, the big ideas are probably still worth it even if you prefer other programs.
If you have a background in SEM, the Mplus team’s lectures on Bayesian SEM are quite nice (i.e., Topic 9 statmodel.com/course_materia…). Though they only show Mplus code, the big ideas are probably still worth it even if you prefer other programs.
10/n
For more informal sources, definitely check in on Gelman’s blog statmodeling.stat.columbia.edu, which sometimes has a very active comments section.
For more informal sources, definitely check in on Gelman’s blog statmodeling.stat.columbia.edu, which sometimes has a very active comments section.
11/n
Some other nice blogs to keep an eye on are by @vuorre vuorre.netlify.com, @krstoffr rpsychologist.com, @tjmahr tjmahr.com, and @djnavarro djnavarro.net. I sometimes blog on Bayes, too solomonkurz.netlify.com/post/.
Some other nice blogs to keep an eye on are by @vuorre vuorre.netlify.com, @krstoffr rpsychologist.com, @tjmahr tjmahr.com, and @djnavarro djnavarro.net. I sometimes blog on Bayes, too solomonkurz.netlify.com/post/.
12/n
Though I’m not a raw Stan user, if you’re interested in going that route, do check out @betanalpha’s writing betanalpha.github.io/writing/, which covers both Stan examples and lots of statistical theory.
Though I’m not a raw Stan user, if you’re interested in going that route, do check out @betanalpha’s writing betanalpha.github.io/writing/, which covers both Stan examples and lots of statistical theory.
13/n
For online support, you should defiantly bookmark the Stan forums discourse.mc-stan.org, which has helpful tags for programs like #brms discourse.mc-stan.org/c/interfaces/b… and #rstanarm discourse.mc-stan.org/c/interfaces/r…. Also the Stan prior wiki github.com/stan-dev/stan/….
For online support, you should defiantly bookmark the Stan forums discourse.mc-stan.org, which has helpful tags for programs like #brms discourse.mc-stan.org/c/interfaces/b… and #rstanarm discourse.mc-stan.org/c/interfaces/r…. Also the Stan prior wiki github.com/stan-dev/stan/….
14/n
You can also find a glut of online lectures from folks on the Stan team mc-stan.org/users/document….
You can also find a glut of online lectures from folks on the Stan team mc-stan.org/users/document….
15/n
Concerning software, you can find some great vignettes for the #brms cran.r-project.org/web/packages/b…, #rstanarm cran.r-project.org/web/packages/r…, loo cran.r-project.org/web/packages/l…, #bayesplot cran.r-project.org/web/packages/b…, #tidybayes cran.r-project.org/web/packages/t… packages.
Concerning software, you can find some great vignettes for the #brms cran.r-project.org/web/packages/b…, #rstanarm cran.r-project.org/web/packages/r…, loo cran.r-project.org/web/packages/l…, #bayesplot cran.r-project.org/web/packages/b…, #tidybayes cran.r-project.org/web/packages/t… packages.
16/n
It just occurred to me: For you SEM lovers, the #blavaan package now supports both JAGS and Stan faculty.missouri.edu/~merklee/blava…. Though I haven’t yet used blavaan to fit a Stan model, I’m excited to test it out sometime soon.
It just occurred to me: For you SEM lovers, the #blavaan package now supports both JAGS and Stan faculty.missouri.edu/~merklee/blava…. Though I haven’t yet used blavaan to fit a Stan model, I’m excited to test it out sometime soon.
17.5/n
For more guidance on how to do this, check out this blog on my experiences using GitHub and #bookdown to make my Statistical Rethinking project. solomonkurz.netlify.com/post/how-bookd…
For more guidance on how to do this, check out this blog on my experiences using GitHub and #bookdown to make my Statistical Rethinking project. solomonkurz.netlify.com/post/how-bookd…
