The Jointprob community: changes on the agenda, and an upcoming talk about Bayesian Hierarchical Models

Posted August 9, 2023 by daslu ‐ 4 min read

TL;DR: In the coming weeks and months, the Jointprob community will organize standalone talks about topics in Bayesian Statistics and Probabilistic Modeling. The first one will be in August, by David MacGillivray, about Bayesian Hierarchical Models. ⭐

The Jointprob community was created by Scicloj during the summer of 2022, aspiring to be a space where friends of diverse backgrounds can learn & explore Bayesian statistics and probabilistic modeling.

We began by reading the first half of Statistical Rethinking by Richard McElreath. In 2023, we had a second reading journey: Bayesian Modeling and Computation in Python by Osvaldo A. Martin, Ravin Kumar, and Junpeng Lao.

After going through most of the book, we had a few discussions and decided to make some changes to our agenda, pace, and format.

In this post, we share some of the plans and invite you to join a couple of sessions later this month.

Recent experiences – our second reading journey

As discussed in our previous blog post, our planned reading for the beginning of 2023 had been Bayesian Modeling and Computation in Python.

During the last few months, we went through most of the core chapters of the book, and a few of the specialized topics.

We wish to thank the authors for this enlightening book, especially Ravin Kumar, who has been following our journey and helping with good advice.

Recently, a few of us have been occupied with other commitments, so it seemed like a good time to pause so that everybody could rest and catch up. This was also an opportunity to rethink our agenda, pace, and format.

We certainly have not finished the book – a lot of it is worth revisiting. Some of our near-term explorations will surely involve returning to specific topics in the book, applying them, and learning more of the relevant foundations.

The state of Jointprob

While reading, Jointprob has been maturing as a community.

The mailing list has about 130 members, who have been more or less active in various ways, offline or online, during the last year. Alongside our biweekly regular sessions (in two groups), we have been active in the community chat and in ad-hoc sessions teaming up on projects.

In a sense, Jointprob has an informal organizing team – about half a dozen people – who are continually involved in thinking and planning our joint learning. Anybody is invited to be part of this ongoing process.

It is never too late to join – we will help you catch up. See the section about Joining at the community page.

Near-term plans – standalone sessions

In the coming weeks and months, our relatively regular sessions will focus on standalone topics – we will meet monthly around a topic prepared by one or two of us.

Sometimes, we will have more than one session about the same topic (say, about a week apart), to make things more accessible to different time zones, etc.

The topics will be grounded in examples and real-world applications. They will be self-contained but may assume some background. Some will be introductory, and some more advanced.

See more details and discussion of these plans at the community chat under the how to continue - Summer 2023 and onwards topic thread of the #general stream.

A special talk: Bayesian Hierarchical Models with David MacGillivray

The first session on this new journey will be a talk by David MacGillivray about Hierarchical Models.

David, who has been presenting many of the topics on our reading journeys last year, as well as his own data modeling projects, will teach the topic through an example from the paper Bayesian hierarchical model for the prediction of football results by Gianluca Baio and Marta A. Blangiardo. ⚽

Reimplementing and exploring some of the paper’s methods using PyMC (version 5), David will demonstrate some of the joys, challenges, and practices of Hierarchical models. We will see a little bit of what might go wrong, as well as some common solutions.

Time This session will be repeated twice, to welcome different time zones:

As usual, the lessons of the first session will probably result in further exploration and polish before the second one.

Length The sessions will be 90 minutes long. Some of us may wish to stay afterwards and chat.

Recording Some parts of the sessions are recorded and shared internally in the Zulip chat. Possibly, we will also share one of the sessions publicly.

Assumed Background For this session, we will assume participants have familiarity with probabilistic programming in PyMC. We will also assume familiarity with core ideas of Bayesian Statistics, equivalent to Chapters 1,2,3,4 of the Bayesian Computation book.

Joining If you wish to be added to our calendar events, please refer to the Joining Jointprob form.