DABEST is a package for Data Analysis using Bootstrap-Coupled ESTimation.

Estimation statistics is a simple framework that avoids the pitfalls of significance testing. It uses familiar statistical concepts: means, mean differences, and error bars. More importantly, it focuses on the effect size of one's experiment/intervention, as opposed to a false dichotomy engendered by P values.

An estimation plot has two key features.

1. It presents all datapoints as a swarmplot, which orders each point to display the underlying distribution.

2. It presents the effect size as a bootstrap 95% confidence interval on a separate but aligned axes.

DABEST powers estimationstats.com, allowing everyone access to high-quality estimation plots.

This package is tested on Python 3.5, 3.6, and 3.7. It is highly recommended to download the Anaconda distribution of Python in order to obtain the dependencies easily.

You can install this package via pip.

To install, at the command line run

pip install --upgrade dabest

You can also clone this repo locally.

Then, navigate to the cloned repo in the command line and run

pip install .

import pandas as pd
import dabest

# Load the iris dataset. Requires internet access.
iris = pd.read_csv("https://github.com/mwaskom/seaborn-data/raw/master/iris.csv")

# Load the above data into `dabest`.
iris_dabest = dabest.load(data=iris, x="species", y="petal_width",
                          idx=("setosa", "versicolor", "virginica"))

# Produce a Cumming estimation plot.
iris_dabest.mean_diff.plot();

Please refer to the official tutorial for more useful code snippets.

Moving beyond P values: Everyday data analysis with estimation plots

Joses Ho, Tayfun Tumkaya, Sameer Aryal, Hyungwon Choi, Adam Claridge-Chang

Nature Methods 2019, 1548-7105. 10.1038/s41592-019-0470-3

Paywalled publisher site; Free-to-view PDF

Please report any bugs on the Github issue tracker.

All contributions are welcome; please read the Guidelines for contributing first.

We also have a Code of Conduct to foster an inclusive and productive space.

We would like to thank alpha testers from the Claridge-Chang lab: Sangyu Xu, Xianyuan Zhang, Farhan Mohammad, Jurga Mituzaitė, and Stanislav Ott.

To test DABEST, you will need to install pytest.

Run pytest in the root directory of the source distribution. This runs the test suite in the folder dabest/tests. The test suite will ensure that the bootstrapping functions and the plotting functions perform as expected.

DABEST is also available in R (dabestr) and Matlab (DABEST-Matlab).