How are you going about this ? What's your workflow to build the model, deploy it, make predictions, etc.

The model is implemented in PyMC3 and runs daily as a batch job. The model state and predictions are stored in DynamoDB.

Early each morning this batch job loads the previous day's posterior as priors, fetches the results for the games played that day, and then fits the Bayesian network to the game results to get new posteriors. I make predictions with these posteriors and then write all the results back to DynamoDB.

The frontend is essentially D3, Django, and Zappa.

I wrote a bit about it in a blog post over the summer: http://everettsprojects.com/2020/08/18/modeling-the-nhl-bett...

There are other posts on my blog about this project. There are links directly to them in the GitHub README: https://github.com/evjrob/bayes-bet#associated-blog-posts