- Video (will be watching together): https://www.youtube.com/watch?v=VVQaIkom5D0
- Video: https://www.youtube.com/watch?v=Zi6N3GLUJmw&list=PLDcUM9US4XdPz-KxHM4XHt7uUVGWWVSus&index=9&t=3964s
- Notes: https://rpubs.com/aseyq/rethinking_week9
- Video: https://www.youtube.com/watch?v=rZk2FqX2XnY
- Book: Chapter 8,9
- Moderator of the week: Piero
- Assignment: https://github.com/rmcelreath/stat_rethinking_2023/blob/main/homework/week04.pdf
- Video: https://www.youtube.com/watch?v=1VgYIsANQck
- Book: Chapter 7
- Moderator of the week: None
- Assignment: None
- Video: https://www.youtube.com/watch?v=uanZZLlzKHw
- Book: Chapter 6
- Moderator of the week: Ali
- Assignment: Q3 of https://github.com/rmcelreath/stat_rethinking_2023/blob/main/homework/week03.pdf (1,2,3)
- Moderator solutions: https://rpubs.com/aseyq/rethinking_week6
- Video: https://www.youtube.com/watch?v=mBEA7PKDmiY
- Book: Chapter 5
- Moderator of the week: Ali
- Assingments: We were not ready for it!
- Mod Demo: https://rpubs.com/aseyq/rethinking_week5
- Video: https://www.youtube.com/watch?v=F0N4b7K_iYQ
- Book: Chapter 4 (contd.)
- Moderator of the week: TBA
- Assignment: Q3 of https://github.com/rmcelreath/stat_rethinking_2023/blob/main/homework/week02.pdf
- Mod Solutions: https://rpubs.com/aseyq/rethinking_week4
- Video: https://www.youtube.com/watch?v=tNOu-SEacNU
- Book: Chapter 4 – Geocentric models (half of it, I suppose)
- Moderator of the week: Sergiu
- Assigment: Q1 and Q2 of https://github.com/rmcelreath/stat_rethinking_2023/blob/main/homework/week02.pdf
- Mod Solutions: https://rpubs.com/aseyq/rethinking_week3
- Video: https://www.youtube.com/watch?v=R1vcdhPBlXA
- Slides:https://speakerdeck.com/rmcelreath/statistical-rethinking-2023-lecture-02
- Book: Chapter 2 - Small Worlds and Large Worlds
- Assignment: Q1,Q2,Q3 of https://github.com/rmcelreath/stat_rethinking_2023/blob/main/homework/week01.pdf
- Mod Solutions: https://rpubs.com/aseyq/rethinking_week2
- Video: https://www.youtube.com/watch?v=FdnMWdICdRs
- Slides: https://speakerdeck.com/rmcelreath/statistical-rethinking-2023-lecture-01
- Book: Chapter 1 - The Golem of Prague
- Assignment: None
- Moderator of the week: Ali
- Link for the live text: https://hackmd.io/@aseyq/rJqezZu3s
- We decided to meet on Thursdays at 17.00
- Ali is the moderator for the first meeting.
Hi there, and thanks for your interest. In this study group, we will work on the acclaimed "Statistical Rethinking" book by Richard McElreath. It is a course on Bayesian Statistics in its core. But regardless of the methodogical stance it has, it contains valuable insights into how we think of data and science. In addition to the book, there is also a set of videos made publicly available by McElreath.
First of all, we'll keep each other accountable. And that's the main target. We all have busy schedules, and it is easy to sacrifice self-study hours when something more urgent comes up. So the first thing is to set a fixed schedule to study each chapter. And with a group of friends, it is much easier and fun to follow it. (And if you decide to join, you will get weekly email reminders for your convenience.)
The bare minimum will be to watch a lecture video on your own once a week. Most of the videos are between 1h and 1h20mins. Ideally one would read the corresponding chapter of the book too, again individually. Then discuss the chapter, ask questions if there is any,and maybe solve some of the questions together at the end of each chapter.
As the name of the group suggests, I imagined and some suggested that we'd gather around aperitivo time, so we could eat and drink at the same time. Some people suggested that it should be during working hours as they have other commitments after work (or have a life of their own. congrats). Probably we can find a sweet spot to accomodate both camps. We thought ~35 minutes for these meetings would be ideal.
We start on the 9th of February.
Based on the videos, there will be 20 "sessions". The original course is 10 weeks. But since we are busy people, I'd asssume watching one video per week, rather than two. IT would be a more managable commitment. But I assume we can go through some chanpters together/faster. So I'd expect to finish all between 16 to 20 weeks, i.e. 4 to 5 months (wicked math skills). Of course there might be some breaks due to holidays etc.
The book defines its target audience as:
The principal audience is researchers in the natural and social sciences, whether new PhD students or seasoned professionals, who have had a basic course on regression but nevertheless remain uneasy about statistical modeling.
I haven't really read the book completely but I suspect some parts might be too obvious and easy for some while some parts might be challenging for most. Nevertheless, my impression is, it is very valuable for anyone who works with data. So if you think you are one of those people, that's great! For the audience, we advertised it initally to a small number of friends to keep the size of the group managable. But feel free to refer friends as long as they commit to the meetings and the rules.
The book is this one:
McElreath, R. (2020). Statistical rethinking: A Bayesian course with examples in R and Stan. Chapman and Hall/CRC.
Here is some more information about it: https://xcelab.net/rm/statistical-rethinking/
It is an expensive book. If you are able to buy it, you can. But feel free to send me an e-mail. I should have some extra copies of it. ;-)
I adapted Richard McElreath's 10 week course and spread it over 20 sessions.
| Session | Meeting date | Reading | Video |
|---|---|---|---|
| 01 | 9 Feb | Chapter 1 | [1] <The Golem of Prague> |
| 02 | 16 Feb | Chapters 2 and 3 | [2] <[Bayesian Inference]> |
| 03 | 23 Feb | Chapter 4 | [3] <[Basic Regression]> |
| 04 | 2 Mar | Chapter 5 | [4] <[Categories & Curves]> |
| 05 | 9 Mar | Chapter 6 | [5] <[Elemental Confounds]> |
| 06 | 15 Mar | Chapter 6 | [6] <[Good & Bad Controls] > |
| 07 | 23 Mar | Chapters 7, 8 | [7] <[Overfitting]> |
| 08 | 29 Mar | Chapter 9 | [8] <[Markov chain Monte Carlo] > |
| 09 | Chapters 10 and 11 | [9] <[Logistic and Binomial GLMs] > | |
| 10 | Chapter 11 | [10] <[Sensitivity and Poisson GLMs] > | |
| 11 | Chapter 12 | [11] <[Ordered Categories]> | |
| 12 | Chapter 13 | [12] <[Multilevel Models]> | |
| 13 | Chapter 13 | [13] <[Multi-Multilevel Models]> | |
| 14 | Chapter 14 | [14] <[Correlated varying effects] > | |
| 15 | Chapter 14 | [15] <[Social Networks]> | |
| 16 | Chapter 14 | [16] <[Gaussian Processes]> | |
| 17 | Chapter 15 | [17] <[Measurement Error]> | |
| 18 | Chapter 15 | [18] <[Missing Data]> | |
| 19 | Chapter 16 | [19] <[Beyond GLMs]> | |
| 20 | Chapter 17 | [20] <[Horoscopes]> |
Of course I can imagine the environment is going to be very friendly. But there will be a number of people involved, and in order to not to waste others' time, we discused some ground rules. Of coure they are likely to change organically as we go further. And feel free to suggest additions/modifications.
Here is a draft of the rules:
1. Punctual Start and Stop: We would start right at the determined time. We won't be able to wait for anyone to join or spend time on chit-chatting after the scheduled time. Similarly we would stop at the given time in order to not to discourage those who have a very limited time. But of course you can join late and catch-up by yourself. Moreover, of course and you can show up early for a little friendly chit-chat, or stay after the sessions.
2. No off-topic conversations: With a friendly atmosphere, it is tempting to talk about other stuff or throw some jokes in to the material. (The videos contain a lot of them!) But during the scheduled time, we should focus on the material. After the session would the a proper time for an off-topic discussions.
3. Self-commitment: It is all about commitment, but if you didn't have time to watch the relevant video or the chapter, of course you are welcomed for the session. But if you are asking a question, or clarification about the material, you are expected to have worked out on the question on your own before.
4. Have fun: Even though we have to have a rigid structure, in the end its about fun. And curiosity and learning is a part of fun. But you can enhance your intellectual fun with some food and drinks.
We though the structure would be as the following:
-
Questions and Discussion (~10 minutes): Here we'd answer each others' questions and discuss the materials.
-
R Code (~18 minutes): The book contains a lot of R codes as demonstration and as questions. We can go through some of them and we can code together. I (ali) am volunteering to share my screen and live code it on the go. But if anyone volunteers for some weeks, that would be more than welcomed.
-
Problem set (~15 minutes): We select some questions from the problem set each week to solve it together.
-
Plan the following week (~2 minutes): At the end of each session, one person can volunteer to be the moderator for the following week. Moderator would be responsible for: (1) selecting a few interesting questions from the set (2) send reminder emails (incl. the question numbers) (3) take care of the timing of the session (4) solve them in the problem session (or attemp). Note that unless you volunteer for it, you wouldn't be moderator and no pressure!
We believe that this would be a fun and rewarding journey. Looking forward to meet you!