Inaugural Book Club
May 27, 2022
A couple of days ago I thought a stats book club would be great but wasn't sure if anyone else thought so too so thought I'd ask the twitterverse.
Thinking of starting a stats book club where each week or two we walk through a chapter, test out models, and generally chat about— RavinKumar (@canyon289) May 27, 2022
Probably over livestream as that's been working out great
Anyone interested in this idea? Ive already got one book in mind
Well my prior uncertainty has been overwhelmed with data. Let's make it happen. For everyone who responded, thank you. You made this happen. If you're certain you want to sign up the mailing list link is below.
Goal, Structure, and Pacing
The goal of this is to encourage continuous learning and sharing of knowledge in an encouraging environment. It's also important that this is accessible to anyone interested. Folks should be able to participate when it works for them, at the pace that works for them.
From the responses its clear no one timezone or day works, and even within a timezone day people are different levels of "busy".
So this is going to be an async first club There'll be Discourse for ongoing community conversation and recorded YouTube streams that I'll host for synchronous sessions (at least for the first book). These two platforms capture discussions and knowledge in a persistent way that balances the magic of synchronous conversations with the async first principle that's guiding this group.
For a working cadence I'm thinking about a two to three week cycle and where we cover a chapter each cycle. As we read we can discuss on Discourse so folks can come in on their own time and catch up on previous conversations. In each cycle I'll also run a livestream where that chapter is discussed and share thoughts and impressions from my personal experience with a question and answer component. The livestream will be recorded keeping with the async first principle. Without actual data this feels like the right balance between "casual walk" versus "all out sprint", synchronous versus async, and "leader led" versus open "community". I'll figure out other ways to make this useful, while sticking to the async first principle, and always with the aim of maximizing the amount of learning.
Are these choices the perfect ones? I'm honestly not sure but let's try it and adjust from there.
The first book: Causal Inference Mixtape
Causal Inference is a topic that's been on my mind for a while. The underlying concepts are universal and common, determine what leads to what, but the skillset to make that determination is far less common.
I'm choosing Causal Inference Mixtape in particular because it has lots of code examples in Python and R, covers a wide range of practical models with real world context, and is very well written.
Most importantly its open access making it accessible to anyone interested though if you have the means you should buy a copy. Purchasing a copy helps the publisher and author continue to put great work out in the open.
How to join
Add your email address to the mailing list using the form below.
I need to get all the technology first and send out an email with further details. The second best is to follow me on Twitter. You don't need to wait though, if you have ideas send me a message.