Seamless A/B testing with Abby.
Project description
abby
A/B testing for Human
Abby is a A/B testing library package for human. Abby aims to make A/B testing as easy as ABC and accessible to anyone.
Installation
$ pip install abby
Quick Start
Please note that your variant name column should be named as variant_name
. Otherwise, the method will raise an error and ask you to adjust the column name accordingly.
A/B testing for continuous metric
from abby.datasets import load_dataset
from abby import compare_ttest
data = load_dataset('click_impression')
compare_ttest(data=data, variants=['control', 'experiment'], numerator='click')
A/B testing for ratio metric
from abby.datasets import load_dataset
from abby import compare_delta
data = load_dataset("click_impression")
compare_delta(
data=data,
variants=["control", "experiment"],
numerator="click",
denominator="impression",
)
A/B testing for ratio metric using bootstrap
from abby.datasets import load_dataset
from abby import compare_bootstrap_delta
data = load_dataset("click_impression")
compare_bootstrap_delta(
data=data,
variants=["control", "experiment"],
numerator="click",
denominator="impression",
n_bootstrap=10_000,
)
A/B testing for multiple metrics
from abby.datasets import load_dataset
from abby import compare_multiple, Ratio
data = load_dataset("click_impression")
result = compare_multiple(
data=data,
variants=["control", "experiment"],
metrics=["click", Ratio("click", "impression")],
)
A/B/N testing for multiple metrics
from abby.datasets import load_dataset
from abby import compare_multiple, Ratio
data = load_dataset("click_impression_3_variants")
result = compare_multiple(
data=data,
variants=["control", "experiment_A", "experiment_B"],
metrics=["click", Ratio("click", "impression")],
)
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
abby-0.1.2.tar.gz
(14.5 kB
view hashes)
Built Distribution
abby-0.1.2-py3-none-any.whl
(14.9 kB
view hashes)