data-utilities

A data analysis and visualization helper module.


Keywords
data, sciences, data-science, utility-library
Licenses
GPL-2.0/GPL-3.0+
Install
pip install data-utilities==1.2.10

Documentation

Build Status

Data Utilities

This module provides some helper functions and conveniences for working with data analysis in python.

It depends on:

  • Numpy
  • Scipy
  • Pandas
  • Matplotlib
  • Seaborn
  • Scikit-learn

Organization and files

.
├── data_utilities
│   ├── __init__.py
│   ├── matplotlib_utilities.py
│   ├── pandas_utilities.py
│   ├── python_utilities.py
│   ├── sklearn_utilities
│   │   ├── grid_search.py
│   │   └── __init__.py
│   └── tests
│       ├── __init__.py
│       ├── test_matplotlib_utilities.py
│       ├── test_pandas_utilities.py
│       ├── test_python_utilities.py
│       ├── test_sklearn_utilities.py
│       └── test_support.py
├── LICENSE
├── MANIFEST.in
├── readme.md
└── setup.py

Each of python's significant data modules has its own set of helper functions.

This module does not intend to create its own API or standards. Instead each of the utilities module should follow the guidelines and APIs provided by the parent module.

Note: This is a primitive project. Expect backwards incompatible changes as I figure out the best way to to develop the utilities.

What's new

  • Added sklearn_utilities.
  • Improved tests customization in du.test.
  • Greatly improved documentation to matplotlib_utilities.
  • Greatly expanded pandas_utilities functions.
  • Improved tests as a whole.

Development guidelines

  • Coding style: PEP 8 compliant.

  • Docstrings: google docstrings.

  • Before commiting new versions do a test for different versions of python3:

    • python3.4
    • python3.5
    • python3.6
    • (newer versions)
    • Rationale: even though stability is expected between python versions some changes occur. See for instance that on commit v1.2.8 (60573d7) there was as unexpected import error on python34 but not on python36.
  • Support first the test interface of numpy:

      `python3 -c "import data_utilities as du; du.test()"`
    

    and then the unittest interface:

      `python3 -m unittest discover -vvv data_utilities/tests`