Welcome to Henchman!

Henchman is a collection of open source python utility functions for working in a jupyter notebook. With Henchman, you can rapidly prototype end-to-end data science workflows. You can explore data with henchman.diagnostics, make interesting plots with henchman.plotting, and do feature selection and machine learning with henchman.selection and henchman.learning.

For more information, visit the Henchman documentation.

Why?

Life is full of reusable functions. Here's what separates Henchman:

• Easy Interactive Plotting: We bypass the flexible Bokeh API in favor of a small, rigid collection of standard data analysis plots. With sliders and checkboxes, finding the right plot parameters can be done with a single function call.

• Memorable API, Extensive documentation: We have a heavy emphasis on ease of use. That means all the functions are sorted into one of 4 semantically named modules and names should be easy to remember inside that module. Additionally, every function has a docstring, an example and a documentation page.

• Novel Functionality: We provide a few functions built from scratch to add to your data science workflow. There are methods to systematically find dataset attributes with overview and warnings from henchman.diagnostics and classes to select features in novel ways with RandomSelect and Dendrogram in henchman.selection.

Install

To install Henchman, run this command in your terminal:

$ python -m pip install fl-henchman

If you are using conda, you can download the most recent build from our channel on Anaconda.org:

$ conda install -c featurelabs henchman

These are the preferred methods to install Henchman, as it will always install the most recent stable release. You can download miniconda from this page.

The sources for Henchman can be downloaded from the Github repo.

You can either clone the public repository:

$ git clone git://github.com/featurelabs/henchman

Or download the tarball:

$ curl  -OL https://github.com/featurelabs/henchman/tarball/master

Once you have a copy of the source, you can install it with:

$ python setup.py install