Bayesian Beta-Binomial Updates: Plotly Dash Application

The code in this repository runs a Plotly Dash app that animates a series of Bayesian updates of beta distributions with binomial data

Application controls:

• Restart Animation - reset the animation to its first state, when only the starting prior beta distribution is displayed
• Play Animation - toggles the animation on or off; 'off' pauses the animation at its current state
• Animation Update Frequency (sec) - the animation updates once every X seconds, where X is chosen here
• Beta parameter alpha starting value (prior) - the starting value for the beta parameter alpha; accepts values with only up to 2 decimal places
• Beta parameter beta starting value (prior) - the starting value for the beta parameter beta; accepts values with only up to 2 decimal places
• Number of binomially distributed data points - how long the sequence of binomial data is, i.e., the number of updates
• Random seed - pseudo-random seed that allows reproducible results whenever the same seed is chosen
• Binomial Proportion (True Beta Distribution Mode) - the proportion used to generate the binomially distributed data

This webpage shows an example of how the beta-binomial model can be used. It also shows videos of beta distribution upates from the Plotly Dash applicaton.

Below is a sequence of images showing a series of beta-binomial updates.

Run with Docker

From the top project directory, you can run these commands:

sudo docker build -t binomial-docker .

sudo docker run -p 8050:8050 binomial-docker

In the commands above, you may substitute a name of your own choosing for binomial-docker.

Then point your web browser at:

http://127.0.0.1:8050/

Run with Poetry

From the 'src' directory, you can run these commands:

poetry install

poetry shell

python beta_binomial.py

Then point your web browser at:

http://127.0.0.1:8050/