BAyesian Model-Building Interface (Bambi) in Python

Bambi is a high-level Bayesian model-building interface written in Python. It works with the probabilistic programming frameworks PyMC3 and is designed to make it extremely easy to fit Bayesian mixed-effects models common in biology, social sciences and other disciplines.

Dependencies

Bambi is tested on Python 3.7+ and depends on ArviZ, formulae, NumPy, pandas, PyMC3 and statsmodels (see requirements.txt for version information).

Installation

Bambi is available from the Python Package Index at https://pypi.org/project/bambi/, alternatively it can be installed using Conda.

PyPI

The latest release of Bambi can be installed using pip:

pip install bambi

Alternatively, if you want the bleeding edge version of the package, you can install from GitHub:

pip install git+https://github.com/bambinos/bambi.git

Conda

If you use Conda, you can also install the latest release of Bambi with the following command:

conda install -c conda-forge bambi

Usage

A simple fixed effects model is shown below as example.

import arviz as az
import bambi as bmb
import pandas as pd

# Read in a tab-delimited file containing our data
data = pd.read_table('my_data.txt', sep='\t')

# Initialize the fixed effects only model
model = bmb.Model('DV ~ IV1 + IV2', data)

# Fit the model using 1000 on each of 4 chains
results = model.fit(draws=1000, chains=4)

# Use ArviZ to plot the results
az.plot_trace(results)

# Key summary and diagnostic info on the model parameters
az.summary(results)

For a more in-depth introduction to Bambi see our Quickstart or our set of example notebooks.

Citation

If you use Bambi and want to cite it please use

Here is the citation in BibTeX format

@misc{capretto2020,
    title={Bambi: A simple interface for fitting Bayesian linear models in Python},
    author={Tomás Capretto and Camen Piho and Ravin Kumar and Jacob Westfall and Tal Yarkoni and Osvaldo A. Martin},
    year={2020},
    eprint={2012.10754},
    archivePrefix={arXiv},
    primaryClass={stat.CO}
}

Contributing

We welcome contributions from interested individuals or groups! For information about contributing to Bambi, check out our instructions, policies, and guidelines here.

Contributors

See the GitHub contributor page.

Contents

• Getting Started
  • Quickstart
  • User Guide
    • Setup
    • Creating a model
    • Data format
  • Formula-based specification
    • Coding of categorical variables
    • Fitting the model
  • Specifying priors
  • Generalized linear mixed models
  • Families
  • Results
  • Plotting
  • Summarizing
  • Accessing back-end objects
• Examples
  • T-test
    • Generate data
      • Model 1
      • Analysis
      • Model 2
      • Analysis
  • Multiple Regression
    • Load and examine Eugene-Springfield community sample data
    • Specify model and examine priors
    • Examine the model results
    • Summarize effects on partial correlation scale
    • Relative importance: Which predictors have the strongest effects (defined in terms of squared partial correlation?
  • Posterior Predictive
  • Robust Linear Regression
  • Logistic Regression
    • Load and examine American National Election Studies (ANES) data
    • Build clinton_model
      • Logistic regression
      • Specify and fit model in Bambi
  • Model assessment
    • Separation plot
    • \(\hat \kappa\) parameter
    • Run Inference
    • Spaghetti plot of model predictions
  • Multi-level Regression
    • Load data
    • Model
    • Estimated regression line
  • Hierarchical linear regression: The sleepstudy example
    • Load data
    • Explore data
    • The model
    • Analyze results
  • Bayesian Workflow (Strack RRR Analysis Replication)
    • Reading in the data
    • Reshaping the data
    • Fitting the model
    • Inspecting results
    • Expanding the model
    • And the answer is…
  • Wald and Gamma Regression (Australian insurance claims 2004-2005)
    • Load and examine Vehicle insurance data
    • Wald family
    • Gamma family
    • Model fit
      • Wald
      • Gamma
    • Model comparison
  • Negative Binomial Regression
    • Negative binomial distribution review
    • Negative binomial in GLM
    • Load and explore Students data
    • Models
      • Model 1
      • Model 2
    • Model fit
      • Model 1
      • Model 2
    • Explore models
    • Plot predicted mean response
  • Binary response: Alternative link functions
    • Generalized linear models for binary response
    • Load data
    • Build the models
      • Logit link
      • Probit link
      • Cloglog link
    • Results
  • Hierarchical Logistic regression with Binomial family
    • Problem description
    • Load libraries and data
    • Explore our data
    • Non-hierarchical model
    • Hierarchical model
    • Compare predictions
    • Footnotes
  • Splines
    • Load Cherry Blossom data
    • Explore the data
    • Determine knots
    • The model
    • Fit model
    • Analisys of the results
    • Plot predictions and credible bands
    • Advanced: Watch out the underlying design matrix
    • Final comments
  • Logistic Regression and Model Comparison with Bambi and ArviZ
    • Adults dataset
    • The model
    • Model 1:
    • Model 2
    • Model 3
    • Model comparison
    • Probability estimation
      • Final remarks
• API Reference
  • bambi.models
  • bambi.priors
  • bambi.families
  • bambi.data
• Frequently Asked Questions
  • Installation issues
    • Bambi not working on Windows
    • Bambi not working on Linux
  • Miscellaneous how to

Indices

• Index

• Module Index

next

Getting Started