Skip to content

sharkutilities/pandas-wizard

Repository files navigation

favicon
pandas-wizard

Documentation Status GitHub Issues GitHub Forks GitHub Stars LICENSE File PyPI - Downloads PyPI Latest Release

GuardRails badge

Pandas-Wizard (pandaswizard) is a simple Python module for providing utility functions and wrappers for the pandas module. The module is kept simple and use of external dependencies is minimized unless needed to enhance performance.

This is a relatively new repository, and if you find any performance or improvement scope please check the contributing guidelines for the organization. All help and criticism are appreciated. If you find any additional use cases please create a pull request or submit for a new feature.

Getting Started

The source code is currently hosted at GitHub: sharkutilities/pandas-wizard. The binary installers for the latest release are available at the Python Package Index (PyPI).

pip install -U pandas-wizard

The list of changes between each release is available here.

The purpose of the below guide is to illustrate the main features of pandas-wizard and assume the working knowledge of the pandas module and use cases. The below example calculates the percentile of pandas.DataFrameGroupBy object using np.percentile.

import pandaswizard as pdw # attempt to create an ubiquitous naming

# let's calculate the 50th-percentile, i.e. the median for each group
percentiles = df.groupby("group").agg({"A" : pdw.percentile(50)})
percentiles.head()

# or, preferred usage is to use in conjunture with other aggregation function like
statistics = df.groupby("group").agg({"A" : [sum, pdw.percentile(50), pdw.quantile(0.95)]})
statistics.head()

The above function calculates the 50th percentile, i.e., the median of the feature "A" based on the grouped column "group" from the data frame.


Footnote: The favicon is designed from the original pandas logo and no copyright infringement is intended. Since the main objective is to provide a utility function for pandas the logo is re-used and developed using canva.