Bayesian Model Evaluation and Criticism
Good statisticians are able to explain their choices, justify their numbers, evaluate their own models, and share their results (in a reproducible fashion)! This tutorial demonstrates how to do all the above, using ArviZ.
Getting things setup
To get started, first identify whether you:
- Prefer to use the
condapackage manager (which ships with the Anaconda distribution of Python),
- Prefer to use
- Prefer to use
- Do not want to mess around with dev-ops, or
- Only want to view the website version of the notebooks.
1. Clone the repository locally
In your terminal, use
git to clone the repository locally.
git clone email@example.com:arviz-devs/bayesian-model-evaluation.git
Alternatively, you can download the zip file of the repository at the top of the main page of the repository. If you prefer not to use git or don't have experience with it, this a good option.
2. Download Anaconda (if you haven't already)
If you do not already have the Anaconda distribution of Python 3, go get it (note: you can also set up your project environment w/out Anaconda using
pip to install the required packages; however Anaconda is great for data science and we encourage you to use it).
3. Set up your environment
If this is the first time you're setting up your compute environment, use the
conda to create an environment.
conda create -n bayes-eval
To activate the environment, use the
conda activate command.
conda activate bayes-eval
If you get an error activating the environment, use the older
source activate command.
source activate bayes-eval
Then follow the instructions for
Please install all of the packages listed in the
pip install -r requirements.txt
An image can be built from the root directory of the repository using the command. This will build an image on your computer with all dependencies and environment
Once an image is built a container can be started with the command
In your terminal a URL for the notebook server will be displayed. Copy and paste that into a browser. With that you'll have Jupyter in a container! If you're using docker you can skip step 4.
3d. Don't want to mess with dev-ops
If you don't want to mess around with dev-ops, click the following badge to get a Binder session on which you can compute and write code.
4. Open your Jupyter notebook in Jupyter Lab!
In the terminal, navigate to this directory and execute
Navigate to the
1_BayesianWorkflow and open notebook
4a. Want to view static HTML notebooks
If you're interested in only viewing the static HTML versions of the notebooks you can view them on github
We would like to thank the whole Bayes community for being open with learnings and material. For this tutorial in particular we'd like to thank Ari Hartikainen, Osvaldo Martin, and Eric Ma for providing feedback and content.