nightingale 1.3

Python library for simplifying statistical analysis and making it more consistent

Nightingale

I named this package Nightingale in honour of Florence Nightingale, The lady with the data.

Installation

You can use pip to install Nightingale:

pip install nightingale

Usage

Population Proportion

from nightingale import get_sample_size, PopulationProportion, get_z_score

print('z-score for 0.95 confidence:', get_z_score(confidence=0.95))
print('sample size:', get_sample_size(confidence=0.95, error_margin=0.05, population_size=1000))
print('with 10% group proportion:', get_sample_size(confidence=0.95, error_margin=0.05, population_size=1000, group_proportion=0.1))

population_proportion = PopulationProportion(sample_n=239, group_proportion=0.5)
print('error:', population_proportion.get_error(confidence=0.95))

Ordinary Least Squares (OLS)

import pandas as pd
import numpy as np
from nightingale import OrdinaryLeastSquares

data = pd.DataFrame({
    'x': np.random.normal(size=20, scale=5), 
    'y': np.random.normal(size=20, scale=5),
})
data['z'] = data['x'].values + data['y'].values + np.random.normal(size=20, scale=1)
print('data:')
display(data.head())

ols = OrdinaryLeastSquares(data=data, formula='z ~ x + y')
print('ols results:')
display(ols.table)

print('r-squared:', ols.r_squared)
print('adjusted r-squared:', ols.adjusted_r_squared)

ANOVA

References

z-score: https://stackoverflow.com/questions/20864847/probability-to-z-score-and-vice-versa-in-python