akashastrub
commented
1 year ago July: "The case for Bayesian" - talk about Bayesian statistics in general. What are the three main reasons why one should favour Bayesian over traditional, frequentist? How will you incorporate it into your work? Talk about elements in which Bayesian is less favourable e.g. depends on audience - at the end of the day science is nuanced and statistics should just enable informed scientific debate. Example of Bayesian - easier to update model once you get new data. Also useful when you simply cannot do it frequentist e.g. imaging. Opening up the black box of pre-made statistical models. Much more intuitive to most people. Generative elements of it are a huge pro.
EOY: "Statistical Rethinking Course Review" - talk about the course. How did you find it? Who would you recommend it to? What are three key takeaways? What did you dislike? How is it structured - would you restructure it? Entropy, MCMC. Nice because super high level learnings that are applicable across all disciplines e.g. entropy.
Q3: "Posit Academy Review"
TBD: "expensifyR"
- Part 1 - the why - targeting potential users that are less technical.
- Part 2 - the how - targeting more technical people who want to know how I did it and with what tools (this can be broken down further e.g. model for classification, Shiny, etc.
MLOps book