Bayesian Modeling Journey (PyMC)
This repository documents my ongoing journey learning Bayesian modeling using PyMC. It contains a collection of models, experiments, and small exploratory notebooks built to understand probabilistic thinking, model construction, and Bayesian inference end-to-end. The focus is on learning and intuition-building, not polished applications. Some models are simple, some are exploratory, and others gradually build toward more complex Bayesian workflows.
- Small Bayesian models and toy examples
- Regression, classification, and generative models
- Monte Carlo simulations and probabilistic intuition
- Experiments with priors, likelihoods, and posterior predictive checks
- Notes and visualizations to understand model behavior
- synthetic or real-data based
- exploratory or structured
- incomplete or iteratively refined
That’s intentional.
- PyMC – probabilistic programming and inference
- ArviZ – posterior analysis and diagnostics
- NumPy / Pandas – data handling
- Matplotlib – visualization
- Google Colab - reproducible execution
The repository is organized loosely by topic or learning stage. File names and notebooks are added as new concepts are explored.
Examples:
- Monte Carlo simulations
- Bayesian linear regression
- Hierarchical models
- Discrete latent-variable models
- Posterior predictive checks
- Model diagnostics and intuition experiments
Most notebooks can be run directly in Google Colab or locally.
To run locally: pip install pymc arviz numpy pandas matplotlib
- Develop intuition for Bayesian modeling
- Understand how priors influence inference
- Learn to diagnose and validate models
- Move from simple examples to structured probabilistic models
- Become comfortable reading, writing, and debugging PyMC models
This is a learning repository. Models are not guaranteed to be optimal, correct, or production-ready. Mistakes, revisions, and experiments are part of the process.