Learning Causal Impact

Set of notebooks with various introductions and examples to several topics relating to forecasting and causal inference for the applied practitioner, including state space models with statsmodels (and pymc_experimental), bayesian inference with numpyro, and an evaluation of various time series utilities (i.e. orbit). Also compares some libraries around causal inference (i.e. CausalPy, tfcausalimpact)

My aim is to have a simple onramp onto Bayesian Structural Time Series by understanding components like:

• Concepts
  • Bayesian Inference
  • Bayesian Workflow
  • Hierarchical Models
  • State Space Models
  • Causal Inference
• Computation
  • jax
  • Modeling and Inference in numpyro and pymc
  • arviz
  • Time Series utilities - causalimpact, orbit, nixtla
  • Wrangling multidimensional data with xarray