Greetings! This repository houses the materials for a presentation given to the Murray Lab on using Markov Chain Monte Carlo for fitting statistical models. The original presentation was on May 3, 2022.
The slides are in the format of a Jupyter notebook written to be presentation-friendly. It can be found at ./mcmc-talk.ipynb and may be updated from time to time. For those interested, the slides were rendered using the Jupyter extension RISE. A few supporting images were pulled from the internet or various texts. Wherever used in the notebook, attribution is given adjacent to the embedded images.
You are highly encouraged to follow along and experiment by running and modifying the Python notebook for yourself. The only required libraries are bambi==0.7.1 (link) and seaborn; many other supporting libraries (numpy, pandas, matplotlib, pymc3, etc.) are pulled in as dependences. An optional but helpful library is graphviz (link), which is used to provide nice visualizations of Bayesian graphical models.
If reading the materials as a static document, the notebook is also available in pdf form at ./mcmc-talk.pdf.
Please feel free to reach out to me with questions/follow-ups, or to access a recording of the original presentation!
Continue to the notebook for more...