Stan is freedom-respecting open-source software for facilitating statistical inference at the frontiers of applied statistics.

• Write-your-own sampler
• Purpose built routines for individual models
  • MCMCPack R package
    
  • Stata's Bayesian package
  • Sawtooth CBC/HB
• Tools which automatically build samplers for a wide class of models
  • WinBugs
    
  • JAGS
    
  • Stan

• Ideal tool for rapid prototyping of new models
• Modern, thoughful software design
• Increasingly attractive option for fitting even standard Bayesian models

• Module 1: Linear regression model
  • Get up-and-running with Stan
    
• Module 2: Multinomial logit model
  • Learn basic Stan syntax
    
  • Test a model using synthetic data
    
• Module 3: Hierarchial multinomial logit model
  • Learn more Stan syntax
    
  • Work with hierarchical models
    
• Module 4: Nested hierarchical multinomial logit
  • Illustrate development of new models using Stan
    
• Bonus: Data fusion for split questionares

Our focus here is on becoming a capable user of Stan, so we won't be spending much time talking about how Stan works internally or how it compares to other Bayesian software or classical inference.

• Kevin Van Horn Senior Data Science Engineer at Adobe
  • Develops commercial-quality software for Bayesian estimation and experimental design
• Elea McDonnell Feit Assistant Professor of Marketing at Drexel University and Senior Fellow of the Wharton Customer Analtyics Initiative
  • Develops tutorials for practitioners in hierarchical Bayes, R, advertising measurement, and field experiments

• R statistical language
  • Manages data input, inspection and cleaning
  • Stores and summarizes posterior draws
  • Provides graphics
• R studio (optional)
  • Convenient user interface for R
• Rstan package
  • Modeling language
  • Statistical inference engine

Within the folder for each module, you will find:
- .html: slides for the module
- .R: R code for the module, so you can follow along
- .csv: data file we will use for estimation

There are other files in each module folder, which we will explain later.